researchspace.auckland.ac - Semantic Scholar...Dr. Sanjeewa Anuruddha Seneviratne A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy in Surgery,

http://researchspace.auckland.ac.nz

ResearchSpace@Auckland

Copyright Statement The digital copy of this thesis is protected by the Copyright Act 1994 (New Zealand). This thesis may be consulted by you, provided you comply with the provisions of the Act and the following conditions of use:

• Any use you make of these documents or images must be for research or private study purposes only, and you may not make them available to any other person.

• Authors control the copyright of their thesis. You will recognise the author's right to be identified as the author of this thesis, and due acknowledgement will be made to the author where appropriate.

• You will obtain the author's permission before publishing any material from their thesis.

To request permissions please use the Feedback form on our webpage. http://researchspace.auckland.ac.nz/feedback

General copyright and disclaimer In addition to the above conditions, authors give their consent for the digital copy of their work to be used subject to the conditions specified on the Library Thesis Consent Form and Deposit Licence.

http://researchspace.auckland.ac.nz/

http://researchspace.auckland.ac.nz/feedback

http://researchspace.auckland.ac.nz/docs/uoa-docs/thesisconsent.pdf

http://researchspace.auckland.ac.nz/docs/uoa-docs/thesisconsent.pdf

http://researchspace.auckland.ac.nz/docs/uoa-docs/depositlicence.htm

Ethnic differences in breast cancer

outcomes in Aotearoa New Zealand

Dr. Sanjeewa Anuruddha Seneviratne

A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

in Surgery, The University of Auckland, 2015.

Abstract

i

Abstract

Background:

Indigenous Māori women experience significantly worse breast cancer outcomes compared

with European women in New Zealand. Underlying reasons for this disparity are complex, and

inadequately explained in the existing literature. This study was aimed at estimating the

Māori-NZ European breast cancer survival disparity, and to identify and quantify impacts of

various factors contributing to this disparity.

Methods:

Data for all women with newly diagnosed invasive breast cancer in the Waikato District

Health Board area between 01/01/1999 and 31/12/2012 were obtained from the Waikato

Breast Cancer Register, and through a retrospective patient clinical notes review. Patient,

tumour and treatment characteristics of Māori and NZ European women were compared in

adjusted multivariable models. Cancer specific survivals were compared using Kaplan-Meier

survival curves, while contributions of different factors towards the survival disparity were

quantified with Cox proportional hazard modelling.

Results:

Of the total of 2791 women included, 2260 (80.1%) were NZ European and 419 (15%) were

Māori. Compared with NZ European women, Māori were significantly more likely to be

diagnosed with more advanced breast cancer, to have comorbidities and to experience longer

treatment delays. Māori were significantly less likely to be diagnosed through screening, to

receive adjuvant radiotherapy and endocrine therapy based on recommended guidelines, and

to be optimally adherent with endocrine therapy.

Compared with NZ European women, Māori had a significantly higher age adjusted cancer

specific mortality (HR=2.02, 95% CI, 1.59-2.58) with significantly lower 5-year (86.8% vs.

76.1%, p<0.001) and 10-year (79.9% vs. 66.9%, p<0.001%) crude cancer-specific survival

rates. Stage at diagnosis explained approximately 40% of the survival disparity while

Ethnic differences in breast cancer outcomes in Aotearoa New Zealand

ii

screening, treatment and patient factors (i.e. comorbidity, obesity and smoking) contributed by

approximately 15% each. The final model accounted for almost all the cancer survival

disparity between Māori and NZ European women (HR=1.07, 95% CI, 0.80-1.44).

Conclusions:

Māori women with breast cancer are twice as likely as NZ European women to die from their

cancer. Lower screening coverage, delay in diagnosis, inferior quality of treatment and greater

patient comorbidity were largely responsible for this survival disparity. Improving healthcare

access and provision of an equitable cancer care for Māori needs greater attention.

iii

Acknowledgements

It is with immense gratitude that I acknowledge the support and help of my supervisors

Professor Ross Lawrenson and Associate Professor Ian Campbell. Their scholarly advice,

meticulous scrutiny and scientific approach have helped me to a great extent to accomplish

this research project.

I am also indebted to Dr. Nina Scott, Māori advisor for her support and advice during this

project. Her vision, enthusiasm and guidance were immensely helpful in me completing this

task and also in approaching the topic in culturally sensitive manner.

I also wish to thank Jenni Scarlet, Rachel Shirley and rest of the staff at the Waikato Breast

cancer Research office who have been supporting and looking after me over the last three

years.

Breast surgeons and other clinical staff at the Waikato hospital and the Breast Care Centre

needs a special thank for all their support and help with my clinical training which also

enabled me to enhance my understanding of practical difficulties and concerns faced by

women with breast cancer in New Zealand.

I sincerely thank the New Zealand Commonwealth Foundation for their Scholarship support

which covered my doctoral fees, and provided me with a stipend.

I wish to thank Professor Nandadeva Samarasekera, Professor of Surgery at the University of

Colombo, for encouraging me to pursue a doctorate, and for providing me with continuous

encouragement and support.

Lastly, and most of all I wish to thank my parents, my wife Sumudu and our kids Senuka and

Oneli for their love and support throughout this journey.


iv

Table of contents

v

Table of Contents

CHAPTER 1. INTRODUCTION ............................................................................................. 1

CHAPTER 2. BACKGROUND – PEOPLE AND HEALTH SERVICES IN

AOTEAROA NEW ZEALAND AND THE WAIKATO REGION ...................................... 5

2.1. People of New Zealand: .................................................................................................... 5

2.2. Māori, British colonization, the Treaty and evolution into present day New Zealand ..... 6

2.2.1 Māori ........................................................................................................................... 6

2.2.2 British colonization ..................................................................................................... 6

2.2.3 Treaty of Waitangi (Te Triti o Waitangi) .................................................................... 7

2.2.4 Present day New Zealand ............................................................................................ 8

2.3. Health care system in New Zealand .................................................................................. 9

2.3.1 Health care structure, organization and delivery ......................................................... 9

2.4. Waikato Region and the Waikato District Health Board ................................................ 13

2.5. Breast Cancer .................................................................................................................. 15

2.5.1 Breast cancer ............................................................................................................. 15

2.5.2 Management of breast cancer .................................................................................... 15

2.5.3 Management guidelines ............................................................................................. 18

CHAPTER 3. LITERATURE REVIEW ............................................................................... 23

3.1. Ethnic inequalities in breast cancer and breast cancer survival ...................................... 23

3.1.1 Ethnic inequalities in breast cancer and breast cancer survival in New Zealand ...... 24

3.1.2 Ethnic inequalities in breast cancer and breast cancer survival in other countries ... 29

3.2. Reasons for ethnic disparities in breast cancer outcomes ............................................... 33

3.2.1 Patient level factors ................................................................................................... 33

3.2.2 Tumour factors .......................................................................................................... 43

3.2.3 Healthcare service factors .......................................................................................... 45

3.3. Understanding key drivers behind ethnic inequities in breast cancer outcomes ............. 54

3.3.1 Healthcare structure (System level)........................................................................... 55

3.3.2 Physician level factors (Provider level) ..................................................................... 58

3.3.3 Patient level factors ................................................................................................... 60

3.4. Conceptual framework .................................................................................................... 62


vi

CHAPTER 4. DESIGN AND METHODS ............................................................................ 65

4.1. Study population ............................................................................................................. 65

4.2. Data sources: ................................................................................................................... 66

4.2.1 The Waikato Breast Cancer Register: ....................................................................... 66

4.2.2 Retrospective data collection: ................................................................................... 66

4.2.3 Other data sources: .................................................................................................... 67

4.3. Data collection: ............................................................................................................... 69

4.3.1 Ethics approval: ......................................................................................................... 69

4.3.2 The Waikato Breast Cancer Register ........................................................................ 69

4.3.3 Retrospective data collection: ................................................................................... 70

4.3.4 Data preparation: ....................................................................................................... 70

4.4. Variables: ........................................................................................................................ 71

4.4.1 Exposure variable - Ethnicity: ................................................................................... 71

4.4.2 Tumour characteristics .............................................................................................. 77

4.4.3 Patient characteristics: ............................................................................................... 81

4.4.4 Health care access ..................................................................................................... 83

4.5. Outcome data .................................................................................................................. 87

4.6. Sample size estimation .................................................................................................... 88

4.7. Data analyses................................................................................................................... 88

CHAPTER 5. RESULTS......................................................................................................... 89

5.1. How valid are the data used in this study? ...................................................................... 91

5.2. What risk factors contribute to ethnic inequities in breast cancer? A preliminary

analysis ................................................................................................................................. 105

5.3. Is breast cancer screening contributing to ethnic inequity in outcomes? ...................... 115

5.4. Are there ethnic differences in breast cancer biology? ................................................. 131

5.5. Are there ethnic differences in delay in surgical treatment? ......................................... 145

5.6. Are there ethnic differences in delay in chemotherapy and radiation therapy? ............ 161

5.7. Are there differences in the use of adjuvant therapy for breast cancer by ethnicity ..... 175

5.8. Does patient adherence with treatment contribute to inequity? .................................... 189

5.9. Are there ethnic differences in quality of surgical care provided for breast cancer? .... 201

5.10. How things add up: quantitative impact of factors on ethnic inequity ....................... 217

Table of contents

vii

CHAPTER 6. DISCUSSION AND CONCLUSIONS ........................................................ 235

6.1. Interpretation of results ................................................................................................. 235

6.1.1 Differences between Māori and NZ European women with breast cancer ............. 236

6.2. Why do these disparities exist? ..................................................................................... 241

6.2.1 Provider and healthcare system characteristics ....................................................... 241

6.2.2 Patient factors including socioeconomic factors ..................................................... 243

6.3. How can we provide Māori women with better and equitable healthcare? .................. 245

6.3.1 Identifying the problem and its underlying reasons ................................................ 245

6.3.2 Providing equitable and acceptable care ................................................................. 245

6.4. Strengths and Limitations – How valid are the findings? ............................................. 252

6.4.1 Study design ............................................................................................................ 252

6.4.2 Internal validity / Bias ............................................................................................. 255

6.4.3 Confounding and estimating effects ........................................................................ 257

6.4.4 Chance and variability ............................................................................................. 260

6.4.5 External validity ...................................................................................................... 260

6.5. Conclusions ................................................................................................................... 263

APPENDICES ........................................................................................................................ 265

REFERENCES....................................................................................................................... 277


viii

List of Tables & Figures

ix

List of Tables

Table 1: Guidelines for timeliness of breast cancer care ........................................................... 20

Table 2: Potential barriers interfering with optimal cancer treatment ....................................... 55

Table 3: Definitions of socio-demographic, tumour and treatment characteristics ................... 73

Table 4: TNM staging system for staging of breast cancer ....................................................... 77

Table 5: SEER summary staging system for staging of invasive breast cancer ........................ 79

Table 6: ICD-9 and ICD-10 codes for breast cancer as underlying cause of death ................... 87

Table 7: Extent of cancer (stage) at diagnosis in the New Zealand Cancer Registry compared

with extent of cancer at diagnosis in the Waikato Breast Cancer Register ............... 96

Table 8: Distribution of characteristics associated with stage known and unknown breast

cancers in the New Zealand Cancer Registry for the Waikato region. ..................... 98

Table 9: Multivariable Cox proportional hazard model for overall mortality risk for unstaged

vs. staged cancer in the New Zealand Cancer Register ........................................... 101

Table 10: Bivariate analysis of risk factors and mortality ....................................................... 109

Table 11: Multivariate model for factors associated with mortality ........................................ 111

Table 12: Multivariate model (conditional logistic regression) for factors associated with

mortality for women with a survival of ≤3 years and >3 years against matched

controls .................................................................................................................... 112

Table 13: Characteristics of women associated with early stage [compared with advanced

stage] at diagnosis of breast cancer by ethnicity for screening age women with

newly diagnosed breast cancer in the Waikato, New Zealand 1999-2012.............. 120

Table 14: Odds ratios for stage at diagnosis (i.e., advanced versus early) in Māori compared

with NZ European women with stepwise adjustment for age, year of diagnosis,

screening status, socioeconomic deprivation and urban/rural residential status ..... 122

Table 15: Adjusted (age and year of diagnosis) breast cancer specific mortality hazard ratios

from Cox regression model ..................................................................................... 123

Table 16: Hazard ratios for breast cancer-specific mortality risk in Māori compared with NZ

European women with stepwise adjustment for age, year of diagnosis, screening

status, socioeconomic deprivation and urban/rural residential status ..................... 125

Table 17: Five-year and 10-year breast cancer specific survival rates by screening status,

ethnicity and socioeconomic deprivation for screening age women with invasive

breast cancer in the Waikato, New Zealand ............................................................ 126

Table 18: Age and breast cancer biological characteristics at diagnosis compared between NZ

European and Māori women. .................................................................................. 136


x

Table 19: Age adjusted odds ratios for tumour biological characteristics by socio-economic

deprivation category (NZDep 2006). ...................................................................... 139

Table 20: Age and deprivation (NZDep 2006) adjusted odds ratios (OR) with 95% confidence

intervals for breast cancer biological characteristics for Māori compared with NZ

European women. .................................................................................................... 139

Table 21: Cox regression model for factors associated with breast cancer specific mortality in

Waikato, New Zealand. ........................................................................................... 140

Table 22: Distribution of age and tumour stage by ethnicity .................................................. 149

Table 23: Delay from diagnosis to primary surgery (in days) by ethnicity ............................. 150

Table 24: Univariate analysis of factors associated with a delay of >31 days and >90 days .. 152

Table 25: Multivariable logistic regression analysis of factors associated with a delay of >31

days and >90 days adjusted for age, stage, deprivation score, mode of diagnosis and

distance from hospital ............................................................................................. 156

Table 26: Univariate analysis of factors associated with delay in first adjuvant therapy, delay

in radiation and delay in chemotherapy for women with newly diagnosed invasive

breast cancer ............................................................................................................ 166

Table 27: Multivariable model for factors associated with delay in first adjuvant therapy, delay

in radiation therapy longer than 90 days and delay in chemotherapy longer than 60

days ......................................................................................................................... 168

Table 28: Cox proportional models for breast cancer specific mortality by delay in first

adjuvant therapy, radiation therapy and chemotherapy .......................................... 170

Table 29: Socio-demographic and tumour characteristics associated with use of adjuvant

chemotherapy for invasive breast cancer in the Waikato, New Zealand ................ 179

Table 30: Multivariable logistic regression analysis for factors associated with use of adjuvant

chemotherapy for invasive breast cancer in the Waikato, New Zealand ................ 181


radiation therapy for invasive breast cancer in the Waikato, New Zealand ............ 182


radiotherapy for invasive breast cancer in the Waikato, New Zealand ................... 184

Table 33: Factors associated with adherence to adjuvant endocrine therapy for hormone

receptor positive invasive breast cancer unadjusted and adjusted multivariable

models ..................................................................................................................... 194

Table 34: Adherence to adjuvant endocrine therapy and breast cancer mortality unadjusted and

adjusted for age, comorbidity, deprivation, tumour factors (size, lymph node status,

grade) and other treatment modalities (surgery, radiotherapy and chemotherapy) 196

Table 35: Adherence to adjuvant endocrine therapy and breast cancer recurrence unadjusted

and adjusted for age, comorbidity, deprivation, tumour factors (size, lymph node

status, grade) and other treatment modalities (surgery, radiotherapy and

chemotherapy) ......................................................................................................... 196


xi

Table 36: Characteristics associated with mastectomy versus breast conserving surgery for

women with T1 & T2 breast cancers undergoing surgery in the Waikato .............. 206

Table 37: Characteristics associated with women undergoing major breast reconstruction

following mastectomy for breast cancer in Waikato ............................................... 209

Table 38: Characteristics associated with women completing definitive local therapy for early

(stage I & II) breast cancer in Waikato ................................................................... 211

Table 39: Patient, tumour treatment and healthcare access characteristics of the study cohort

by Māori and NZ European ethnicity ...................................................................... 222

Table 40: Breast cancer specific mortality hazard ratios for selected variables from final Cox

proportional hazard model ...................................................................................... 226

Table 41: Hazard ratios for breast cancer-specific mortality risk in Māori compared with NZ

European women with stepwise adjustment for demographics, screening status,

disease factors, treatment factors, patient factors and healthcare access ................ 228


xii


xiii

List of Figures

Figure 1: The structure of the New Zealand health system in 2008 .......................................... 11

Figure 2 - Map of Waikato District Health Board catchment area ............................................ 13

Figure 3: Registration rates for female breast cancer in New Zealand, by ethnicity ................. 24

Figure 4: Age standardised breast cancer mortality rates in New Zealand by ethnicity (for ages

1–74 years) ................................................................................................................ 25

Figure 5: Cancer specific survival from breast cancer in Māori and non-Māori, 1996–2001 ... 27

Figure 6: Model of pathways to cancer treatment ..................................................................... 34

Figure 7: Distribution of stage at diagnosis for breast cancer in New Zealand,1996-2001 ...... 35

Figure 8: Cancer detection pathway .......................................................................................... 36

Figure 9: Age standardized mortality rates for breast cancer in New Zealand by deprivation

quintiles ..................................................................................................................... 37

Figure 10: Neighbourhood socioeconomic deprivation (NZDep2006) for Māori and non-Māori

2010 ........................................................................................................................... 38

Figure 11: An illustration of the overall milestones and time intervals in the route from first

symptom until start of treatment ............................................................................... 49

Figure 12: Conceptual framework depicting complex interaction of patient, health system and

cancer related factors influencing clinical outcomes ................................................ 63

Figure 13: New Zealand Statistics – Urban/Rural Classification system .................................. 84

Figure 14: Faster Cancer Treatment Indicators 2012-2013 ....................................................... 86

Figure 15: Distribution of accurately staged, inaccurately staged and unstaged breast cancer in

the New Zealand Cancer Registry compared with the Waikato Breast Cancer

Register...................................................................................................................... 96

Figure 16: Trends in proportional distribution of cancer stage in the Waikato Breast Cancer

Register compared with staged cancers in the New Zealand Cancer Registry. ........ 97

Figure 17: Trends in unstaged, accurately staged and inaccurately staged breast cancer in the

New Zealand Cancer Registry compared with the Waikato Breast Cancer Register99

Figure 18: Kaplan-Meier survival curves by cancer stage for invasive breast cancers included

in the New Zealand Cancer Registry and the Waikato Breast Cancer Register for the

Waikato region ........................................................................................................ 100

Figure 19: Kaplan-Meier curves for breast cancer specific survival for screen detected and

symptomatically detected breast cancers in screening age women by ethnicity..... 126


xiv

Figure 20: Ten-year breast cancer specific survival rates by socioeconomic deprivation (based

on Kaplan-Meier survival curves by socioeconomic deprivation quintile) for

screening age women .............................................................................................. 127

Figure 21: Kaplan-Meier survival curves for non-screen detected cancers in screening age NZ

European and Māori women by socioeconomic deprivation status ........................ 128

Figure 22: Distribution of percentage of women with a delay longer than 31 days by ethnicity

(Māori vs. NZ European) and age category ............................................................ 151

Figure 23: Yearly trend in proportional delay longer than 31 days by ethnicity (Māori vs. NZ

European) ................................................................................................................ 155

Figure 24: Time trends in delay in adjuvant radiation therapy longer than 90 days (Panel A)

and adjuvant chemotherapy longer than 60 days (Panel B) for invasive breast cancer

in Waikato, New Zealand ........................................................................................ 169

Figure 25: Time trends in delay in adjuvant chemotherapy longer than 90 days for invasive

breast cancer in Waikato, New Zealand .................................................................. 169

Figure 26: Time trends in use of chemotherapy and radiotherapy by ethnicity ..................... 185

Figure 27: Annual rates of high level of adherence (MPR≥80%) with adjuvant endocrine

therapy for hormone receptor positive invasive breast cancer for NZ European and

Māori women .......................................................................................................... 193

Figure 28: Kaplan-Meier survival curves for 5-year breast cancer specific and disease free

survival by adherence to adjuvant endocrine therapy in Waikato, New Zealand ... 197

Figure 29: Trends in rate of sentinel lymph node biopsy for women with early stage (stage I &

II), T1-2, cN0 tumours undergoing an axillary surgical intervention by ethnicity . 208

Figure 30: Trends in the rates of post-mastectomy breast reconstruction by ethnicity ........... 211

Figure 31: Kaplan-Meier survival curves for breast cancer specific survival (unadjusted) for

Māori and NZ European cohorts ............................................................................. 225

Figure 32: Age structure for Māori and non-Māori populations in Waikato in 2006 ............. 236

Figure 33: Four possible targets for interventions to reduce ethnic/socioeconomic inequalities

in health ................................................................................................................... 246

Figure 34: Intervention framework to improve health and reduce inequalities ....................... 247

List of publications

xv

List of Publications

1. Seneviratne S, Campbell I, Scott N, Shirley R, Peni T, Lawrenson R (2014) Accuracy

and completeness of the New Zealand Cancer Registry for staging of invasive breast

cancer. Cancer Epidemiology 38: 638-644.

2. Seneviratne S, Campbell I, Scott N, Lawrenson R, Shirley R, Elwood M (2015) Risk

factors associated with mortality from breast cancer in Waikato, New Zealand: A case

control study – Public Health (epub ahead of print)

3. Seneviratne S, Campbell I, Scott N, Shirley R, Lawrenson R (2015) Impact of

mammographic screening on ethnic and socioeconomic inequities in breast cancer

stage at diagnosis and survival in New Zealand: a cohort study. BMC Public Health 15:

46.

4. Seneviratne S, Scott N, Shirley R, Kim B, Lawrenson R, Campbell I (2015) Breast

cancer biology and ethnic disparities in breast cancer mortality in New Zealand: a

cohort study – PLOS ONE (In press)

5. Seneviratne S, Campbell I, Scott N, Coles C, Lawrenson R (2015) Treatment delay for

Māori women with breast cancer in New Zealand. Ethnicity and Health 20: 178-193.

6. Seneviratne S, Campbell I, Scott N, Kuper-Hommel M, Round G, Lawrenson R (2014)

Ethnic differences in timely adjuvant chemotherapy and radiation therapy for breast

cancer in New Zealand: a cohort study. BMC Cancer 14: 839.

7. Seneviratne S, Campbell I, Scott N, Lawrenson R. Ethnic differences in use of

adjuvant therapy for breast cancer in New Zealand. Submitted for publication in

Australian and New Zealand Journal of Public health

8. Seneviratne S, Campbell I, Scott N, Kuper-Hommel M, Kim B, Lawrenson R (2015)

Adherence to adjuvant endocrine therapy: Is it a factor for ethnic differences in breast

cancer outcomes in New Zealand? The Breast 24: 62-67.

9. Seneviratne S, Scott N, Lawrenson R, Campbell I (2015) Ethnic, socio-demographic

and socioeconomic differences in surgical treatment of breast cancer in New Zealand,

ANZ Journal of Surgery (epub ahead of print)

10. Seneviratne S, Campbell I, Scott N, Shirley R, Peni T, Lawrenson R. Ethnic

differences in breast cancer survival in New Zealand: Contributions of differences in

screening, treatment, tumour biology, demographics and comorbidities, Submitted for

publication in Cancer Causes & Control


xvi

Co-authorship forms

xvii

Co-Authorship Forms


xviii

Co-authorship forms

xix


xx

Co-authorship forms

xxi


xxii

Co-authorship forms

xxiii


xxiv

Co-authorship forms

xxv


xxvi

xxvii

Abbreviations

AI – Aromatase inhibitors

ALND – Axillary lymph node dissection

BCS – Breast Conserving Surgery

BCSM- Breast cancer specific mortality

BMI – Body mass index

BSA – BreastScreen Aotearoa

CCC- Cancer care coordinator

CI – Confidence interval

DHB – District Health Board

ER – Oestrogen receptor

FNAC – Fine needle aspiration cytology

FSA- First specialist assessment

HCP - Health Care Provider

HER-2 – Human epidermal growth factor receptor – type 2

HR – Hazard ratio

LN – Lymph node

LVI – Lympho-vascular invasion

MDT – Multidisciplinary team

MPR – Medication possession ratio

MRI – Magnetic resonance imaging

NHI – National Health Index

NMDS – National Minimum Dataset

NZCR – New Zealand Cancer Registry

OR – Odds ratio

PR – Progesterone receptor

SD – Standard deviation

SEER – Surveillance Epidemiology and End Results

SNB – Sentinel lymph node biopsy

TNM – Tumour, Node, Metastasis

US scan – Ultrasound scan

WBCR – Waikato Breast Cancer Register

WHO – World Health Organization


xxviii

Introduction

1

Chapter 1. Introduction

Approximately 3,000 New Zealand women are diagnosed with breast cancer and more than

600 women die each year (1). New Zealand has the seventh highest age standardized mortality

rate from breast cancer in the world; a rate which is 20% higher than Australia (2, 3).

Māori, the Indigenous New Zealand population has the highest known breast cancer incidence

of any population group in the world, and this incidence is 28% higher compared to NZ

European women (117.2 vs. 90.6 per 100,000 population) (1). Māori women have experienced

an increase in breast cancer incidence over the last 20 years compared to a decline in incidence

among NZ European women (1). Māori women have a lower survival rate from breast cancer

which together with the higher incidence translates into a 60% higher breast cancer mortality

compared to NZ European women (31.8 vs. 17.9 per 100,000 population) (4, 5). While there

has been an improvement in breast cancer survival for both Māori and NZ European over last

two decades, a significant gap in survival persists (6).

Factors that contribute toward inequities in breast cancer survival in New Zealand may

include; health service and patient factors including stage at diagnosis, access, timeliness and

quality of cancer treatment and tumour factors including biological behaviour (5, 7-9). Delay

in diagnosis in Māori leading to more advanced breast cancer at diagnosis has repeatedly been

shown to be a major factor for lower survival rates in Māori women (4). However, Māori

women have been shown to have a 32% higher stage adjusted mortality compared to non-

Māori, hence factors other than stage at diagnosis make an important contribution to mortality

inequity (4).

Irrespective of their origins or basis of formations, different ethnic groups in different

countries have been shown to have substantially different diagnosis rates, treatment, and

outcomes for a variety of diseases, including cancer (4, 10, 11). Explaining these differences

has inherent value, in that the understanding and treatment of all patients with breast cancer

may be improved. A clear understanding of the relative contributions of each of the

contributing factors toward breast cancer disparities between Māori and NZ European women

will inform future research and guide strategies to address breast cancer inequities.

Achieving equity along cancer care pathways could result in large and rapid gains for cancer

control, especially when compared to developing new treatments or reducing incidence and

help improve outcomes for all.


2

Reducing inequities in cancer mortality using quality improvement measures could

theoretically result in about a 50% reduction in breast cancer mortality inequities seen between

Māori and NZ European women (12). Research is needed to identify where inequities in

access, timeliness and quality of care occur. Such information could inform the development

of quality improvement initiatives to achieve equity along the breast cancer care pathway

between Māori and NZ European women.

Most data for New Zealand research into breast cancer inequities have been from National

Datasets (13, 14). Although the New Zealand National Cancer Registry is a valuable resource

for population based research into cancer, it has significant limitations for researching causes

of inequities due to inconsistent recording of disease stage at diagnosis and limited data on

treatment (15, 16). Currently available data alone have been insufficient to fully understand

ethnic inequities in breast cancer outcomes in New Zealand and its underlying causes.

This thesis examines differences in socio-demographic, tumour and treatment characteristics

and their impacts on breast cancer survival disparity between Māori and NZ European women.

It also seeks to explore possible reasons for why these factors may differ between Māori and

NZ European women. Data for this thesis come from a cohort of women diagnosed with breast

cancer in the Waikato, New Zealand between 1999 and 2012. It also draws on published

literature on cancer inequities in general, and on breast cancer inequities in specific, which

have provided the platform for the current research.

Cancer is a complex illness; its management even more complex. Cancer control covers the

spectrum from cancer prevention through screening, diagnosis, treatment, rehabilitation to

provision of palliative care. Due to the nature of the data availability, this thesis has restricted

the analyses to elements within secondary and tertiary care; that is from the point of first

specialist assessment onwards. Where feasible, data prior to the first specialist assessment,

including breast cancer screening data have also been included. As the focus of this thesis is

breast cancer outcomes in relation to mortality and recurrence, rehabilitative and palliative

care services are not examined.

Introduction

3

The specific objectives of this thesis are as follows:

1. To determine differences in access to care factors (i.e., breast cancer screening,

socioeconomic status, ethnicity and residence) on stage at diagnosis

2. To determine the association of age, residence, ethnicity, public/private treatment and

screen/non-screen presentation on delay and/or quality of care (i.e., delay in primary

surgery, type of surgical treatment, usage and delay in adjuvant therapy and adherence

with adjuvant therapy)

3. To determine ethnic differences in cancer biological characteristics (i.e., histology,

tumour grade, lympho-vascular invasion, oestrogen/progesterone receptor [ER/PR]

status, HER-2 status)

4. To determine ethnic differences in breast cancer outcomes and factors contributing to

such differences

To determine the magnitude of breast cancer specific survival disparity between

Māori and NZ European women, stratified by stage at diagnosis.

To determine differences in survival for different treatment groups and different

biological markers

To determine the quantitative impact of these factors on the overall ethnic

disparity

Chapters of this thesis and their purposes are as follows:

Chapter 1- Introduction: Provides an introduction to the topic of the thesis and defines the

specific objectives. It also clarifies the focus of this thesis.

Chapter 2 – Background: This chapter provides an insight into the context in which breast

cancer disparities between Māori and NZ European populations are generated. It includes a

history of New Zealand, political and health system changes leading to the current political

landscape and healthcare system. The last section of this chapter includes a brief description of

breast cancer and it basic management principles.


4

Chapter 3 - Literature Review: This chapter is aimed at reviewing available literature on

cancer disparities, and underlying causes, both locally and internationally. It attempts to

critically analyse literature to identify existing gaps, to build a platform and create a niche for

this research project. It also attempts to analyse different theories on ethnic disparities to

identify key drivers and finally attempts to build a conceptual framework on which the study

analyses are based upon.

Chapter 4 - Design and Methods: This chapter includes a detailed description on how the

study was conducted including study sample selection, data collection, data definitions and

data analyses.

Chapter 5 – Results: This chapter is set out in ten sections; each a separate study aimed at

clarifying a specific area in relation to breast cancer ethnic disparity. Each study is set up with

a brief introduction, specific methods used for the study, results and a focussed discussion of

findings.

Chapter 6 – Discussion: This chapter is kept relatively brief as most of the study results and

their specific meanings and implications are discussed under each section in Results. The

discussion attempts to bring together and summarize all of the study findings and then

attempts to fill the gaps identified during the literature review to create a clearer picture on

why ethnic disparities might occur and what strategies are available to tackle these disparities.

This chapter will also critically analyse strengths and limitations of the study, and finishes

with a conclusions section which summarizes the whole thesis.

Background

5

Chapter 2. Background – People and health services in Aotearoa

New Zealand and the Waikato region

2.1. People of New Zealand:

Māori are the Indigenous people of New Zealand. They are Polynesian people who settled in

New Zealand about 1000 years ago. Subsequent immigrations, initially by Europeans starting

from 1800’s and later by Pacific Islanders and Asians, have created the present day dynamic

multicultural society of New Zealand. According to the latest population census in 2013, NZ

Europeans (Pākehā) make the numeric majority (74%), while Māori comprise approximately

15% of the population. Asian (11.8%) and Pacific (7.4%) comprise the other two main ethnic

groups (17).

Māori were the sole inhabitants of Aotearoa New Zealand for several hundred years until the

arrival of the British and subsequent colonization which started in the early 1800s. Through a

combination of colonization, the Treaty of Waitangi and military suppression of Māori

resistance, the British settlers managed to make New Zealand a part of the British Empire (18).

Although the Treaty granted Māori equal rights, acquisition of Māori land by the British led to

the New Zealand land wars in 1860’s. Following these wars, over 100,000 hectares of Māori

land was confiscated by the government under the New Zealand Settlements Act in 1863,

purportedly as punishment for Māori rebellion. Loss of land, illnesses and after effects of

tribal wars led to a rapid decline in the Māori population with many scientists and politicians

predicting an imminent extinction in the late 1800s (19).

A major resurgence in the Māori population was observed after the Second World War. Many

Māori migrated to larger towns and cities during the depression and post-world war periods in

search of employment, leaving rural communities depleted and disconnecting many urban

Māori from their traditional ways of life (20). While standards of living improved among

Māori during this time, they continued to lag behind Pākehā in areas such as health, income,

skilled employment and access to higher education. Māori leaders and government

policymakers alike have struggled to deal with social issues stemming from increased urban

migration, including a shortage of housing and jobs, and a rise in crime, poverty and health

problems.


6

2.2. Indigenous Māori, British colonization, the Treaty and evolution into

present day New Zealand

2.2.1 Māori

Māori are the Indigenous people in NZ and are considered tangata whenua –people of the

land. They are descendants of Polynesians who had arrived approximately 1000 years ago (18,

21).

Iwi (tribes) formed the largest social groups of Māori who trace their ancestry to original

Polynesian migrants. Each iwi comprised of a structure for ruling and leadership was based on

a system of chieftainship. Important units of pre-European Māori society were the whānau or

extended family, and the hapū or group of whānau. After these came the iwi or tribe,

consisting of groups of hapū. Traditional Māori society preserved history orally through

narratives, songs, and chants; skilled experts could recite the tribal genealogies (whakapapa)

back for hundreds of years.

Traditional Māori belief systems, such as views about reliance on the whānau, individual mana

(prestige), death and dying, and practices associated with tapu (sacred), continue to influence

health behaviour. These views may influence preferences for care, individual help-seeking

behaviour and responses to health care providers (22).

2.2.2 British colonization

The first significant contact between Māori and Europeans occurred in 1769, at the time of

James Cook’s expedition to New Zealand from Britain (23). Subsequently, by early 1800s

many Europeans traders, missionaries, convicts escaped from Australia, and deserting seamen

began to settle in the country.

From about 1800 to about 1830, the Europeans lived in New Zealand on Māori sufferance.

During this period a rapid decline in Māori population was observed partly because thousands

were killed in their musket wars in the 1820s and 1830s, but mainly because the Europeans

brought a battery of new germs and viruses. Measles, some forms of influenza, typhoid, small

pox and other diseases reduced the Māori population from perhaps 200,000 in 1769, when

James Cook arrived, to less than 40,000 by 1870.

http://www.teara.govt.nz/glossary#tangata whenua

Background

7

With the growing British population in New Zealand and due to the threat from the French, the

British Crown intended to annex New Zealand into the British Empire. The authority of New

Zealand was taken by the British Crown through the signing of the Treaty of Waitangi in 1840

(24).

2.2.3 Treaty of Waitangi (Te Triti o Waitangi)

In 1840 the Treaty of Waitangi, a formal agreement for British settlement and a guarantee of

protection of Māori interests, was signed by representatives of the British crown and Māori

chiefs (24). This is considered as the foundational document of New Zealand as a modern

state. The Treaty exists in both Māori and English texts with important differences between

the two (25).

Māori and colonial British had different expectations of the Treaty in accordance with the

language, worldview and political agenda of each. Māori chiefs expected to retain self-

governance at tribal level largely unaltered with the Treaty (26). However, the British regarded

the Treaty as a mechanism through which to create a relationship where “the Crown as

sovereign and the Māori as subject” (24, 25). The Crown’s interpretation of the Treaty

prevailed despite the objections of Māori chiefs who signed the Treaty (24, 25).

It is estimated that the Māori population numbered approximately 80,000 at the time of

signing the Treaty, along with a population of about 2,000 settlers. Signing of the Waitangi

treaty facilitated a large-scale influx of British migrants. Rapid influx of settler Europeans and

a decline in Māori resulted in a major change in country’s demographics. By 1901, the settler

population of 770,000 outnumbered the Māori population by 16.5:1 (23).

Majority of the Māori regarded the Treaty as the basis for their relationship with the British

Crown and sought to have their rights as the Indigenous people addressed (27). The British

Crown considered the content of the Treaty to be largely irrelevant and considered it only as a

document through which the British established its ‘ownership’ of New Zealand (27).

Since the 1970s, public awareness of the Treaty of Waitangi has continued to increase,

primarily as a result of growing Māori aspirations for self-determination. In recent government

health documents, the Indigenous status of Māori has been recognized, and the Treaty of

Waitangi has been acknowledged as a fundamental component of the relationship between

Māori and the government (28, 29). However, the Treaty has never been included in social


8

policy legislation, and there is a clear gap between acceptance of the Treaty and translation of

its aims into actual health gains for Māori (30).

The Treaty’s primary intention was to protect and maintain the well-being of all citizens,

although many of its components were not implemented by the colonial government. The

Treaty has relevance for health care services and how these should address needs of Māori

communities to have equitable care (31). Further, the Treaty supports the principle of tino

rangatiratanga (self-determination) that encourages the development of Māori providers where

Māori deliver health care for Māori. (32). Despite the provisions in the Treaty and subsequent

policies supporting equitable care, Māori continue to experience poor health compared with

Pākehā populations (31).

2.2.4 Present day New Zealand

The present day New Zealand landscape reflects the relationships and interactions of Māori

and European settlers over 200 years. Māori continue to be disadvantaged with poor

educational achievements, high rates of unemployment and poor health compared with NZ

Europeans (33-35). Implementation of several policies specifically aimed at improving

education, employment and health among Māori have seen gradual improvements that have

narrowed the disparity between Māori and Europeans. Still, lack of an organized, integrated

approach towards correcting these differences seems to be deficient.

Several new policies were implemented especially over the last two decades with increased

government spending on Māori specific programmes (36). These efforts have seen some

improvements in health and education for Māori, although some of these efforts have been

criticised for lack of benefit in relation to the amount of money spent, due to poor planning

and implementation. Absence of a broad, integrated approach for Māori development and

implementation of programmes which were piecemeal or sector specific are also believed to

be responsible for not achieving expected results out of these programmes.

Background

9

2.3. Health care system in New Zealand

The healthcare system of New Zealand has been primarily a public funded system since

introduction of the Social Security Act in 1938. It has undergone significant structural changes

throughout past several decades (37). The Labour Government introduced the Social Security

Act in 1938 with the objective of providing free and universally available healthcare for all.

However, the British Medical Association (i.e., the organisation of doctors) campaigned

against a salaried system, as it interfered in the relationship between doctors and their patients.

The Government compromised by instituting a General Medical Services benefit, but doctors

reserved the right to set their fees and to charge an additional fee. Hence, ‘private practice’

survived the 1938 reforms.

By 1970s the rapid growth in health costs resulted in a series of changes being introduced to

improve efficiency and health outcomes. Although the public health system, especially the

hospital side had developed rapidly, there was still a strong private sector especially for

surgical interventions.

2.3.1 Health care structure, organization and delivery

The health care system in New Zealand comprises an interlinking system of primary,

secondary and tertiary health services. From its establishment in 1930’s the evolution of the

health system has resulted in various changes.

The health system that was established based on the 1938 reforms included minimal regard to

needs or aspirations of the Māori society. Nonetheless, universal coverage remained one of the

prime goals of this health structure, which aimed to provide accessible healthcare for all

citizens, including Māori. Primary care remained within the private sector though it was

subsidized by the government (37). Some of the primary characteristics of this initially laid

down health system remain in the present New Zealand health system including subsidization

of primary health care and universal publicly funded hospital care (37).

In 1987, the Area Health Board Act created 14 Area Health Boards in place of Hospital

Boards, which existed following the 1938 Social Security Act. Later, in 1993 the Health and

Disability Services Act was introduced creating 23 Crown Health Authorities that functioned

under four Regional Health Authorities. The current District Health Board (DHB) system was


10

introduced in 2001 creating 21 DHBs (reduced to 20 in 2010) to deliver public health care in

New Zealand.

During the evolution of the health system, it was dominated by Pākehā needs and beliefs with

minimal input or involvement of Māori. First Māori involvement in health care organization

was reported in mid 1980s; even then it was more symbolic than with active participation in

decision making (37). The more recent health service changes have seen a greater participation

of Māori in health care organizations and health care decision making processes. Despite that,

under-representation of Māori in health organisations, lack of focus on equity and health

delivery processes remains a significant concern in addressing Māori health issues.

The majority of the burden for core healthcare system expenditure rests with government

(approximately 77% in 2005). Private payment by individuals also plays an important role in

the overall system, although the costs of these payments are comparatively minor. In 2009,

New Zealand spent 10.3% of gross domestic product (GDP) on health care compared with

8.7% in Australia, 9.8% in the United Kingdom and 17.4% in the USA. However, per capita

health expenditure in Australia (US $ 4179) and USA (US $ 7163) were much greater than

that of New Zealand (US $ 2917) (38).

The current structure of the health system has a greater devolution of service planning and

funding and, greater emphasis on primary health care services (Figure 1). Twenty DHB’s are

responsible for funding and provision of services at a local level. All public hospitals come

under the care of DHBs, who are responsible for providing secondary and tertiary health care.

Primary health care (provided by general practitioners) and community based services are also

funded by DHBs, but are delivered by independent providers (36).

The current health system places more emphasis on Māori health compared with previous

manifestations (36). DHBs are required to have at least two Māori members which need to be

increased based on population structure of the DHB area. DHBs are specifically required to

work towards reducing health disparities by improving health outcomes for Māori and other

disadvantaged population groups. Further, they are required to establish and maintain

processes to enable Māori to participate in and contribute to strategies for Māori health

improvement.

Private sector health care in New Zealand is mostly provided at secondary care level. These

institutions are focused on providing specialist elective services which include elective

surgical and other invasive procedures. Patients are referred to private hospitals or to

Background

11

appropriate specialist consultants by general practitioners. All surgical procedures for breast

cancer including major breast reconstructions are provided through private hospitals. However

the provision of oncology services is limited in the private sector. Although specialist

outpatient care is freely available, private parenteral chemotherapy facilities are available only

in a few areas such as Auckland and Palmerston North. Radiotherapy facilities have been

recently started in the private sector in Auckland.

Figure 1: The structure of the New Zealand health system in 2008 (Source - Ministry of Health)

National Cancer Control Strategy

The National Cancer Control Strategy was initiated in 2003 with the two key objectives of

reducing the incidence and impact of cancer, and reducing inequalities with respect to cancer

(21). As with many government documents, the National Cancer Control Strategy also refers

to three principles derived from the Treaty of Waitangi; partnership (between Māori and

Crown agencies), participation (of Māori in service planning and delivery) and protection (of

Māori health and health equity with non-Māori). Through its action plan the Cancer control


12

Strategy has put forward an intervention framework to improve health outcomes and reduce

inequalities (21). It specifically targets to reduce the burden of cancer by making structural

changes in the society through economic and social policy changes, and by improving access

and care pathways in health and disability services.

BreastScreen Aotearoa

BreastScreen Aotearoa (BSA) is the New Zealand National Breast Screening Programme, a

free mammographic breast screening service available for all screening age women. Since it

was established in 1999, the BSA has provided free biennial mammographic screening for all

women aged between of 50 to 64 years, and this age range was extended to include women

aged 45 to 49 and 65 to 69 years from July 2004.

Mammographic screening coverage in New Zealand has gradually picked up over the last

decade and has achieved the target biennial coverage of 70% for NZ European women since

2010 (39). However, poor screening coverage has remained a significant issue for Māori

women for whom the coverage was only 62.7% in 2012 (39). There was a large variability in

screening coverage rate for Māori by region, which ranged from 54% to 79% across the

country in 2012 (40). There is close monitoring of coverage by region and by ethnicity, and

targets are set yearly with screening providers to improve these rates.

Background

13

2.4. Waikato Region and the Waikato District Health Board

The Waikato region which is situated in the central part of North Island of New Zealand, is

based on the catchment of the Waikato River and includes Lake Taupo and Rotorua regions.

This is similar to the catchment area of the Waikato Hospital Board formed in 1938. The

Waikato Area Health Board (formed in 1989) was based on a smaller area, excluding Taupo

and Rotorua. Boundaries of the Waikato Area Health Board have largely remained unchanged

during the creation of Waikato Crown Health Authority in 1993 and more recently the

Waikato District Health Board in 2001 (Figure 2).

With a population of 365,000 Waikato District Health Board (DHB) is the fourth largest in

New Zealand. The Waikato DHB covers an area of 21,220 square kilometres and stretches

from northern Coromandel to close to Mount Ruapehu in the south and from Raglan on the

west coast to Waihi on the east. It has a major urban centre (i.e., Hamilton), a significant rural

population and a Māori population of more than 76,000 (21% of the Waikato DHB

population) which is the second highest in New Zealand. Waikato Māori comprise 13.5% of

the total New Zealand Māori population (41).

Figure 2 - Map of the Waikato District Health Board catchment area


14

Breast cancer services in the Waikato DHB area are provided in the public sector through

specialist services located at the tertiary hospital in Hamilton. In addition, surgical treatment is

also provided through several well equipped private hospitals. Radiation therapy services for

the Waikato region are provided exclusively through a radiation facility at the tertiary hospital

in Hamilton, while chemotherapy facilities are provided through a satellite site (i.e., Thames)

in addition to the tertiary hospital in Hamilton.

Background

15

2.5. Breast Cancer

The following section provides a brief overview of breast cancer and basic principles of

management.

2.5.1 Breast cancer

Breast cancer primarily includes malignant tumours arising from epithelial cells of ducts and

lobules of the breast. Other types of breast cancers include sarcomas and phyllodes tumours

which originate, behave and are treated differently, and hence have been excluded from this

study.

Breast cancer is the commonest cancer among women in the world with 1.7 million new breast

cancers diagnosed in 2012, which represents about 12% of all new cancer cases and 25% of all

cancers in women (42). Incidence of breast cancer is substantially higher in developed

countries compared with less developed countries. New Zealand, with an age standardised

incidence of 93.3 per 100,000 population in 2008, was ranked 7th highest in the world (43).

2.5.2 Management of breast cancer

Most early breast cancers are curable while metastatic breast cancer is considered incurable.

Surgery is considered as the primary treatment modality for all early breast cancers. Adjuvant

therapy is recommended based on future risk of local and/or systemic failure. Advanced

breast cancers (non-metastatic) are managed with neo-adjuvant therapy followed by surgery

while metastatic cancer is managed mostly with primary systemic therapy.

Investigations and diagnosis

Breast cancers are diagnosed through two main methods; through mammographic breast

cancer screening and through non-screen (symptomatic) presentations. In New Zealand,

approximately a third of breast cancers overall, and approximately two thirds of cancers within

screening age are diagnosed through mammographic screening.

A breast lump is the most common (>80%) presenting symptom for symptomatic breast

cancer. Other symptoms include breast pain, nipple discharge, axillary or neck lump/s or

symptoms of metastatic disease. A woman with symptoms suspicious of a breast cancer is

initially assessed by a primary care physician / general practitioner and is referred to a

specialist breast care unit for further management.


16

Triple assessment forms the cornerstone of breast cancer diagnosis. This includes clinical,

radiological and pathological assessment aimed to diagnose or exclude cancer in all women

presenting with symptoms suggestive or suspicious of a breast cancer. Standard imaging

includes mammogram and ultra sound scan (US scan) of the breast while Magnetic Resonance

Imaging (MRI) is reserved for situations including assessment for multi-focality in lobular

cancer or for women with mammographically dense breasts.

Once clinical and imaging assessments are completed, a pathological examination is done to

complete the triple assessment. Traditionally fine needle aspiration cytology (FNAC) was

used, but this has gradually been replaced with core biopsy as the preferred modality. Core

biopsy has a higher sensitivity and provides better histologic detail which helps in decisions,

including for neo-adjuvant therapy. The majority of core biopsies are performed under image

guidance (US or stereotactic) further increasing the diagnostic yield of a core biopsy. Other

pathological techniques for diagnosis include punch biopsy (when the cancer involves the

skin) or open excisional biopsy that is performed in situations where triple assessment fails to

exclude or confirm a cancer.

Following confirmation of the diagnosis of breast cancer, further management depends on

several of patient and tumour characteristics. Clinical tumour stage which encompasses both

clinical and image findings and patient fitness for treatment are the most important of these

factors. If the patient is relatively young and fit, cancer stage (early or advanced) alone

governs further treatment. In general terms, early breast cancers (stages I & II) are treated with

primary surgery while advanced cancers (stage III) are treated with neo-adjuvant therapy

(chemotherapy or endocrine therapy) which is aimed at down-staging the cancer and to make

it operable. Metastatic cancers are managed with the objectives of controlling the disease to

prolong life and to manage symptoms, as these are deemed incurable.

Surgery

Early breast cancers are treated with primary surgery if the patient is fit to undergo surgery.

Primary surgery is aimed at completely removing the primary tumour and removing involved

axillary nodes or staging the axilla. Mastectomy (removal of the whole breast) and breast

conserving surgery (BCS) are the two main surgical options for the breast. Axilla is treated

with axillary lymph node dissection (ALND) if there is clinical evidence of node involvement.

If the axilla is clinically normal, it is managed based on a sentinel node biopsy (SNB). SNB

Background

17

attempts to identify the first node/s draining the axilla and if it is positive for tumour cells on a

frozen section, an ALND is done. If SNB is free of tumour, no further surgery is performed on

the axilla. Reconstruction of the ipsilateral breast (after mastectomy) using either autologous

tissue or an implant or a combination, constitutes the final step in surgical treatment.

Radiation therapy

Radiotherapy forms another loco-regional treatment modality which is aimed at reducing the

risk of local recurrence of breast cancer in the preserved breast after BCS or in the chest wall

after mastectomy. All women after a BCS are required to receive radiation therapy while after

a mastectomy the need depends on the risk of local recurrence. For instance, New Zealand

breast cancer guidelines recommend radiation therapy for all women if ≥4 lymph nodes are

involved or if the primary tumour diameter is ≥50mm (44). External beam radiation is

provided through a linear accelerator and a woman is given 40-50Gy of radiation delivered

over 15 to 25 fractions.

Chemotherapy

Chemotherapy is one of the systemic breast cancer treatments, and together with endocrine

therapy has been responsible for the major reduction in breast cancer mortality observed over

the last two to three decades. Chemotherapy is aimed at targeting and killing any residual

tumour cells remaining after primary surgery. Chemotherapy is highly toxic to rapidly

dividing cells that include cancer cells. Due to its toxicity, side effects and lack of action on

slow growing cancers, chemotherapy is generally reserved only for aggressive or advanced

cancers which are generally associated with poor outcomes. However, selection of women

who would get a significant survival benefit (generally considered as >5%) is complicated.

Although several algorithms and software programmes (e.g. Adjuvant online!, Predict) help in

this decision (45, 46), still they lack the necessary precision. Newer tumour genome based

tests including OncotypeDX® and Mammaprint® have improved this precision to predict

which women would benefit from chemotherapy (47, 48), but for a significant minority the

dilemma of chemotherapy remains, even with these tests.


18

More effective chemotherapy regimens have replaced the traditional CMF

(cyclophosphamide, methotrexate and 5-fluorouracil) regimen and now include doxorubicin

and cyclophosphamide followed by a taxane (e.g., docetaxel).

Hormonal / Endocrine therapy

A majority of breast cancers are dependent on female hormones (i.e. oestrogen and

progesterone) for their growth and sustenance. Blockage or removal of this hormonal

stimulation is a highly effective strategy of treating breast cancer in women whose cancers are

hormone receptor positive. Traditionally, tamoxifen, a selective oestrogen receptor modulator

(SERM) has been used, but has largely been replaced by aromatase inhibitors for post-

menopausal women (49). Still, tamoxifen is the preferred agent for pre-menopausal women.

Traditionally, hormonal therapy was prescribed for five years. However, recent trials have

shown additional mortality benefit of longer term tamoxifen up to 10 years as compared with

the conventional 5-year therapy (50, 51).

Biological therapy

Amplification of human epidermal growth factor receptor type-2 (HER-2) in breast cancer is

associated with aggressive cancer behaviour and poor outcomes. Blockage of HER-2 with

trastuzumab (brand name - Herceptin®) has helped to significantly reduce the risk of breast

cancer mortality in these women.

2.5.3 Management guidelines

Guidelines for the management of early breast cancer were first published in New Zealand in

2009 (44). Prior to this, management of breast cancer in New Zealand was based on guidelines

published in other countries including Australia and the United Kingdom (52-54).

Staging

All breast cancer patients are expected undergo clinical staging at the time of diagnosis which

include assessment of clinical tumour size, loco-regional node involvement and a general

Background

19

physical examination along with a history to check for possible symptoms of distant disease

(55). Several staging investigations may also be used, including chest X-ray, bone

scintigraphy, liver imaging, computerised tomography (CT), positron emission tomography

(PET), and serum biomarkers (e.g. CA 15-3).

Staging investigations are generally reserved for women with features suggestive of advanced

breast cancer or specific symptoms that may indicate presence of metastatic disease (44).

Staging investigations are generally performed prior to surgery, however may also be

performed after primary surgery in situations where a woman is found to have more advanced

disease than expected after pathological assessment of the resected specimen.

Multidisciplinary care

Breast cancer care provided through a multidisciplinary breast team (MDT) has been shown to

be an effective means of establishing a correct diagnosis in women referred with breast

symptoms and has also been established to be an efficient, cost-effective way to care for

women with breast cancer (56). MDTs also provide the opportunity for useful second

opinions. Cases of all New Zealand women diagnosed with breast cancer are expected to be

discussed at least once at a MDT, that should comprise of Surgeons, Pathologists, Radiologists

and Oncologists (Medical and Radiation) (55). However, this is not compulsory and hence

cases of some women with breast cancer diagnosed and managed at secondary care institutions

and private sector are not routinely discussed at MDTs.

Timeliness of treatment

Guidelines on timeliness of diagnosis and treatment were available for women diagnosed

through the BSA programme since 1999 (57) . However, no guidelines on timeliness were

available in New Zealand for women with symptomatic breast cancer until the Faster Cancer

Treatment Indicators and the standards of service provision for women with breast cancer

were published in 2012 and 2014, respectively (58, 59). Based on current guidelines, women

with high risk of breast cancer are expected to have their first specialist assessment (FSA)

within 14 days and receive first treatment within 31 days of confirming diagnosis (Table 1)

(58, 59).


20

Table 1: Guidelines for timeliness of breast cancer care (Adapted from: Faster Cancer Treatment

Guidelines 2012 and Standards of Service Provision for Breast Cancer Patients 2014)

Timely access – referral

Women referred urgently with a high suspicion of breast cancer - FSA within 14 days

Women referred with a moderate suspicion of breast cancer - FSA within 30 days

Women referred with a low suspicion of breast cancer - FSA within 90 days

Timely access – primary treatment

Women referred urgently with a high suspicion of breast cancer - first cancer

treatment within 62 days

Women with a confirmed diagnosis of breast cancer - first cancer treatment within 31

days of the decision to treat

Timely access – systemic treatment

Women recommended adjuvant systemic therapy by an MDT and fit to receive it -

commence treatment within six weeks of surgery for breast cancer

Timely access – radiation treatment

Women with breast cancer referred for radiation oncology assessment - FSA with a

radiation oncologist within two weeks of receipt of referral (where chemotherapy is

not part of the management)

Women consenting to radiation therapy after surgery - commence treatment once the

surgical site has healed and within six weeks of surgery (where chemotherapy is not

part of management)

FSA – First Specialist Assessment

Follow up

Follow up after treatment of breast cancer is aimed at early detection of tumour recurrences or

development of new breast cancer. In addition this provides an opportunity for detection and

management of therapy related complications and to provide psychological support for these

Background

21

women (59, 60). In New Zealand setup, an annual clinical examination and mammography are

provided through a specialist breast service for 10 years for all women treated with breast

cancer who are in good health (44).

Changes in management guidelines

Although the general principles of management of breast cancer have remained largely

unchanged, significant changes in specific aspects of management are observed over last 10-

15 years.

In relation to surgical treatment, a major change has been the management of the axilla. All

women with invasive cancer were uniformly treated with an axillary lymph node dissection up

to late 1990s. However, early 2000’s a significant change was observed where more and more

women were offered sentinel lymph node biopsy (SNB) based management for the axilla.

Currently, all women who have early breast cancer with clinically node negative axillae are

offered SNB based management (44, 59).

Other major changes were observed in systemic therapy for breast cancer. Traditional CMF

chemotherapy was gradually replaced with AC (doxorubicin and cyclophosphamide)

chemotherapy in late 1990s, and over last 10 years has seen further changes with the addition

of taxanes (44). Further, trastuzumab became available in early 2000’s, and was made

available through the public health care system in New Zealand for HER-2 amplified early

breast cancer from 2007 (61).

Significant changes in endocrine therapy include the use of aromatase inhibitors as the

preferred endocrine therapy for post-menopausal women in place of traditional tamoxifen and

prolongation of endocrine therapy from conventional five years for up to 10 years (62).

Another major advancement in cancer treatment and especially in treatment of breast cancer

has been the gradual incorporation of personalized cancer treatment. As of now, personalized

treatment for cancer is only beginning, with a small number of validated drug-test companion

products available. Currently, personalized treatment is most advanced for breast cancer, and

tests such as ER, HER-2, Oncotype DX, and MammaPrint are available to personalize

treatment decisions. Although the evolution of personalized treatment is likely to be slow

personalized treatment for every patient with breast cancer is likely to be available in the near

future.


22

Literature Review

23

Chapter 3. Literature review

This chapter describes and discusses evidence on Indigenous non-Indigenous and ethnic

inequalities in breast cancer, breast cancer care and outcomes. This review is aimed at

informing these questions based on current evidence what is the extent of ethnic disparity in

breast cancer survival in New Zealand, and why ethnic disparities in breast cancer outcomes

might occur. The review was performed by combining a formal literature review with a search

of references and bibliographies from a range of books and documents, without applying any

specific inclusion or exclusion criteria.

This chapter starts with a discussion on Indigenous and ethnic disparities in cancer with a

special emphasis on breast cancer, and is followed by a discussion on underlying causes and

theories for these disparities. The chapter concludes with a conceptual framework that is

aimed at understanding the key factors contributing to ethnic disparities in breast cancer.

3.1. Ethnic inequalities in breast cancer and breast cancer survival

Ethnic and geographic differences in cancer including breast cancer are extremely variable,

and are believed to be due to a multiplicity of factors. For instance, geographically, women

from Europe, North America and Oceania have experienced significantly higher incidences of

breast cancer compared with women from Asian and African countries (3, 43). Similar

disparities in breast cancer incidence are observed among different ethnic groups in many

multi-ethnic countries, including the USA, Australia, Canada and New Zealand (63).

Historically, many Indigenous and ethnic minority populations have had low incidences of

cancer compared with majority or non-Indigenous populations of their respective countries

(63). The extent of these differences have been observed to become smaller over successive

generations, as the Indigenous and minority populations have taken up cancer-promoting

environmental exposures, at rates similar to or higher than that of majority/settler populations.

In contrast, many of these Indigenous and minority ethnic groups have continued to

experience far worse cancer survival rates than the majority/non-Indigenous populations of

respective countries. Poorer cancer survival rates combined with increasing cancer incidences

have resulted in a rapidly worsening cancer burden among Indigenous and minority

populations (63).


24

The following section first analyses ethnic disparity in breast cancer and breast cancer survival

in New Zealand. Then a comparison is performed with ethnic disparities in breast cancer from

other countries.

3.1.1 Ethnic inequalities in breast cancer and breast cancer survival in New Zealand

Major ethnic disparities in cancer incidence and outcomes are documented in New Zealand.

Poorer cancer outcomes for Māori compared with non-Māori have been well known for over

three decades (64-66). In 2008, overall cancer incidence in Indigenous Māori was 19% higher

compared with non-Māori (220 vs. 185 per 100,000 age standardized population) (4). The

extent of this disparity is quadrupled for cancer mortality as the mortality rate in Māori was

78% higher compared with non-Māori (112 vs. 63 per 100,000 population) (67).

Approximately a quarter of the Māori non-Māori cancer survival disparity has been shown to

be due to breast cancer (4).

Figure 3: Registration rates for female breast cancer in New Zealand, by ethnicity, 2000–2010 (Source

– New Zealand Cancer Registry)

Note: The rate shown is the age-standardised rate per 100,000 female population, standardised to the WHO

world standard population.

Literature Review

25

Māori women have one of the highest incidences of breast cancer in the world (42). The

incidence of breast cancer in Māori is approximately 30% higher compared with non-Māori

(117.2 vs. 90.6 per 100,000 age standardized population) and appears to be increasing further,

while the incidence in non-Māori is on the decline (1). This has created a rapidly widening gap

in incidence between Māori and non-Māori women (Figure 3). For instance, over the period

between 2000 and 2010, breast cancer incidence in non-Māori has declined by about 5% while

the incidence in Māori has risen by 10-15% (1). The increase in breast cancer incidence in

Māori is largely unexplained at present. Increasing rates of obesity, alcohol consumption and

changes in reproductive behaviours all contribute to the increase, but not fully account for it

(33).

Figure 4: Age standardised breast cancer mortality rates in New Zealand by ethnicity (for ages 1–74

years) (68)

A similar trend is observed for breast cancer mortality where the rates have been gradually

increasing for Māori while a gradual reduction is seen for non-Māori women (Figure 4). Many

factors are believed to be contributing to this inequality in breast cancer mortality between

Māori and non-Māori women (7, 13). Several authors have attempted to identify these factors,

and to quantify the contribution of each of these factors towards Māori non-Māori breast

cancer survival disparity. Almost all such studies to date have used routinely collected

administrative data from the NZCR. Although this approach has provided large numbers of

breast cancers for these studies, high percentages of missing data and possible biases including


26

miscoding and misclassification of the NZCR data have significantly restricted the validity of

these studies (4, 6, 13). Furthermore, absence of data in the NZCR on some of the important

aspects of breast cancer inequalities including cancer treatment and comorbidities has

prevented a complete analysis of factors contributing to ethnic disparities in breast cancer

outcomes in New Zealand. For instance, based on the NZCR data, Māori women with breast

cancer have been shown to be diagnosed with more advanced staged breast cancer compared

with NZ European women (Figure 2). However, after adjusting for stage, breast cancer

mortality rate is still about 40% higher in Māori which indicates a major contributions of other

factors, including treatment differences, towards this disparity (4).

The Unequal Impact report published in 2006 has analysed New Zealand cancer data by

ethnicity for the period from 1996 to 2001 (69). Its second edition Unequal Impact-II included

cancers diagnosed between 2002 and 2006, and overall disparities are reported for the 11 year

period from 1996-2006 (4). Both these reports are based on data from the NZCR and have

included more than 1,000 and 2,000 Māori and more than 10,000 and 20,000 non-Māori breast

cancers in Unequal Impact and Unequal Impact-II, respectively.

Both Unequal Impact reports reported significantly higher incidences and lower breast cancer

survival rates for Māori compared with non-Māori women (Figure 5). Age standardized rate

ratios for breast cancer incidence and mortality have shown marginal increases from the first

to second reports. The incidence ratio has increased from 1.21 (1.14-1.28) to 1.28 (1.20-1.37)

while the morality hazard ratio has increased from 1.68 (1.54-1.88) to 1.73 (1.55-1.94). An

increase in mortality hazard ratio was observed for distant (HR increase from 1.34 to 1.51)

stage cancer, while a substantial reduction in mortality hazard was reported for localized

breast cancer (HR from 1.87 to 0.74) in Māori compared with non-Māori women from the first

to second reports. Despite these variations, the overall stage adjusted breast cancer mortality

hazard ratio has shown only a marginal reduction from the first to second report (HR from

1.48 to 1.44). According to these reports, advanced stage at diagnosis has contributed by

approximately a third to the observed mortality disparity, with only minimal change overtime.

The high proportion of unstaged cancer included in Unequal Impact and Unequal Impact-II

reports (17% and 15% respectively) has impacted on the accuracy of these quantifications,

especially as Māori were over-represented in the unstaged breast cancer category. Despite

these limitations, Unequal Impact reports provide useful information on the extent of Māori

non-Māori breast cancer disparity, and trends in breast cancer incidence and mortality

disparity between Māori and non-Māori women over time.

Literature Review

27

Figure 5: Cancer specific survival from breast cancer in Māori and non-Māori, 1996–2001

(unadjusted) (Source – Unequal Impact – II)

The fourth version of the Hauora - Māori Standards of Health has also performed an analysis,

similar to the Unequal impact using the NZCR data, but only included women diagnosed over

a five-year period between 2000 and 2004 (31). Incidence and mortality hazard ratios

published in this report are much similar to Unequal Impact reports with an incidence rate

ratio of 1.14 and a mortality hazard ratio of 1.71 for Māori compared with non-Māori women.

In 2005, Jeffreys and colleagues published relative 5-year breast cancer survival by ethnicity

using data from the NZCR, for women diagnosed between 1991 and 2004 (9). This study

included a total of over 16,000 breast cancers, and included over a thousand of Māori breast

cancers. Age adjusted 5-year survival rates were 76% and 82% for Māori and non-Māori non-

Pacific women, respectively. Adjusting for stage at diagnosis resulted in almost equal 5-year

relative survival rates of 81% and 82 % for Māori and non-Māori non-Pacific women,

respectively in their analysis. These findings are quite contradictory to Unequal Impact, where

only a third of the survival disparity was observed to be due to stage at diagnosis, despite

using data from the same source. Methodological differences and differences in statistical

methods used in these two studies may have contributed to these variations. For example, the


28

Unequal Impact report has used Cox proportional hazard modelling while the Jefferys study

has used relative survival for comparison of survival between Māori and non-Māori women.

The Cancer Trends in New Zealand for 1991-2004 report compiled by Soeberg and colleagues

analysed relative survival trends for Māori and non-Māori cancer including breast cancer,

adjusting for socioeconomic and income status (6). According to models used in this report,

over the 15-year period from 1991 to 2004, breast cancer mortality for Māori and non-Māori

declined significantly, with a greater decline for Māori than for non-Māori. Although the

excess mortality in Māori compared with non-Māori has declined by approximately 50% over

each 10-year period, even in 2004, mortality rate in Māori was about 20% higher than for non-

Māori women. Similar substantial reductions in mortality for socioeconomic and income

categories were documented in this report. However, in contrast to ethnicity, the mortality gap

between the lowest and the highest socioeconomic groups have remained essentially

unchanged over the 15-year study period.

McKenzie and colleagues have performed a series of studies with the aim of quantifying the

impact of cancer stage, healthcare access and tumour characteristics on breast cancer survival

disparity between Māori and non-Māori women (12, 13, 70). Similar to studies described

previously, these studies also used the NZCR data, but included additional data including

socioeconomic status and tumour biological characteristics in their analyses. Their first study,

which used data for women diagnosed over a 2-year period from April 2005 to April 2007

from the NZCR, reported an unadjusted breast cancer mortality hazard ratio of 1.73 for Māori

compared with non-Māori women (13). Adjusting for age, socioeconomic and tumour

characteristics, completely explained the survival disparity, with a final adjusted hazard ratio

for Māori of 1.03. However, several important characteristics in this study needs clarification.

First, this study included very high percentages of missing data which were over 60% for

some key variables. Data imputation has been used to overcome this issue, but this potentially

has added further bias. For instance, the authors have considered missing data as missing at

random (which it was not, with higher percentages of data missing for Māori and for women

with poor survivals) and have performed a simple multiple imputation (71). A different

imputation method such as multiple imputation combined with either chained equations or

modelling could have been used, which might have helped to overcome the issue of non-

random missing data and provide more accurate estimates (72). The same study has reported

an adjusted hazard ratio of 1.43 for Māori compared with non-Māori, who had complete data,

which differs substantially from final analysis based on data imputation.

Literature Review

29

Further, this study has reported that there were significant differences in breast cancer

biological characteristics between Māori and non-Māori women. However these differences in

biology were not found to be contributing to breast cancer mortality disparity. Limitations of

the study analysis, as discussed above, are likely to have limited the validity of these findings.

Pacific women have a lower breast cancer incidence than Māori and European women in New

Zealand (73). However, a rapid rise in breast cancer incidence among Pacific women has been

observed over last three decades. Compared to NZ European women, Pacific women are more

likely to be younger at diagnosis, present with more advanced disease and have prognostic

phenotypes associated with poor survival (70, 74). Pacific women have a breast cancer

survival rate which is worse than Māori women. A rapidly rising incidence combined with a

lower survival has resulted in a threefold increase in breast cancer mortality among Pacific

women over a two decade period (7). Literature on breast cancer for Pacific women are sparse.

Overall, based on existing literature, there is conclusive evidence on breast cancer incidence

and mortality disparity between Māori and non-Māori women, where Māori have about 20-

30% and 60-70% higher age standardized rates of incidence and mortality, respectively

compared with non-Māori women. Some studies have shown a substantial reduction in

mortality disparity over last 10-20 years while others have shown only a marginal decrease.

Advanced stage at diagnosis in Māori appeared to be the biggest contributor to this disparity

with a reported contribution of approximately 30-40%, although some have reported this to be

close to 100%. Although differences in socioeconomic status and cancer biological

characteristics appeared to be contributing to the survival disparity, its quantitative impact is

unclear at present. None of the studies to date has provided any data on possible inequalities in

treatment or its contribution towards mortality disparity.

3.1.2 Ethnic inequalities in breast cancer and breast cancer survival in other countries

Ethnic disparities in breast cancer and breast cancer outcomes have been reported from many

countries including the USA, Australia, Canada and the UK.

Breast cancer disparities in the USA

Breast cancer disparities in the USA are mostly reported in relation to disparities between

White European and African American women, and less frequently including Hispanic and


30

Asian women. Comparatively, limited data are available on breast cancer disparities between

settler White American women and Indigenous populations in the USA which include

American Indians and Alaskan Natives.

African Americans

In the USA, African American women bear a greater burden of breast cancer, similar to Māori

in New Zealand. Breast cancer mortality rate is about 40% higher in African American than in

White women, despite a 10-20% lower breast cancer incidence (75).

The disparity in breast cancer mortality between African American and White American

women has been known for almost half a century (76). A disparity in stage adjusted breast

cancer survival between African American and White women was first reported in the 1980’s,

from a study based on US national breast cancer statistics (77), which changed the earlier

notion that poorer outcome among African Americans were solely due to advanced stage at

presentation (78). Since then several studies have shown more aggressive tumour biology,

higher rates of comorbidity and inferior quality of breast cancer care being factors contributing

to poorer outcomes in African American compared with White American women in the USA

(79-81).

A large number of studies have examined breast cancer outcome differences between African

American women and White American women, and underlying causes for these differences,

especially over the last decade. Findings from some of the larger studies and meta-analyses are

discussed below.

Newman and colleagues published a meta-analysis of 20 studies published between 1980 and

2005 on ethnic inequalities in breast cancer outcomes in the USA (82). They reported that

African American ethnicity to be an independent predictor for worse overall mortality (HR-

1.27, 95% CI 1.18-1.38) and breast cancer mortality (HR=1.19, 95% CI 1.10-1.29) compared

with White American women, after adjusting for cancer stage, age and socioeconomic status.

This analysis confirmed findings of several previous studies, which have demonstrated more

aggressive cancer biology and inferior quality cancer treatment for African American

compared with White American women as major contributors for breast cancer survival

disparity in the USA (80).

Curtis and colleagues attempted to identify and quantify the impact of different factors

contributing to ethnic disparity in breast cancer survival in the USA, with a sample of over

40,000 women over the age of 68 years, diagnosed with breast cancer between 1994 and 1999

Literature Review

31

(83). In this study, the unadjusted breast cancer mortality rate for African American women

was 63% higher compared with White American women. Adjusting for screening status, stage

at diagnosis, comorbidity, tumour biology and treatment almost completely explained the

survival disparity with a final adjusted hazard ratio of 1.08. Lowest unadjusted and adjusted

breast cancer mortality hazards were observed for Asian and Pacific Island women (HR=0.59

and 0.61, respectively).

Silber and colleagues also investigated reasons for ethnic disparity in breast cancer survival in

the USA by comparing a cohort of African American women with breast cancer over the age

of 65 years with three matched cohorts of White American women, diagnosed during 1991-

2005 (81). The 5-year survival rate for African American women was 12.9% lower than for

White women, and approximately two-thirds of this disparity was due to more advanced stage

at diagnosis in African American women. Most of the residual disparity was explained by

comorbidity and differences in tumour biology. Although significant differences in treatment

including longer delays and lower use of chemotherapy was observed in African American

women, these differences apparently contributed to <1% of the survival disparity. However,

these results are substantially different from several other published studies including a meta-

analysis published by Blackman and colleagues (80, 84). This analysis has reported that

disparities in breast cancer treatment as a substantial and significant contributor towards the

final survival disparity between African and White American women.

Overall, based on published studies, advanced stage at diagnosis appears to be contributing to

approximately a third to a half of the mortality disparity between African American and White

women in the USA, while the rest seems to be distributed among cancer biology, treatment

differences and patient comorbidity.

American Indians and Alaskan Natives

The incidence of breast cancer among Indigenous populations in the USA are substantially

lower compared with White American women (58 vs. 141 per 100,000 age standardized

population in 2000) (85). Comparatively, a much narrower mortality gap (14.9 vs. 27.2 per

100,000 age standardized population in 2000) is observed between Indigenous and White

women due to substantially lower survival rates in Indigenous women. Similar with African

American women, delay in diagnosis, high rate of comorbidity and inferior quality of

treatment are the likely major causes for poor survival in Indigenous women. However, unlike

for African American women, cancer biology has not been shown to be a factor for lower

survival among Indigenous women (86).


32

Breast cancer disparities in Australia

Indigenous Australians including the Aborigines and Torres Straight Islanders have been

documented to have significantly poor health outcomes, including from cancer, compared with

settler European population (87-89). Bramley and colleagues published a comparison of

survival disparities between Indigenous and settler populations in the USA, Canada, Australia

and New Zealand (63). Of these countries, the greatest absolute and relative survival disparity

between Indigenous and settler populations was observed in Australia.

The incidence of breast cancer among Indigenous populations is about half of that of European

women in Australia (89), which is similar to Indigenous American Indians in the USA (85).

Similarly, Aborigines were reported to have worse survival from cancer, as shown in a

matched cohort study of Indigenous and non-Indigenous Australians with cancer by Valery

and colleagues (88). Aboriginal patients were found to be diagnosed with more advanced

cancer, were more likely to have comorbidities and were less likely to receive surgery,

chemotherapy and radiotherapy for treatment of cancer. The worse hazard ratio for mortality

persisted even after adjusting for stage, comorbidity and treatment differences indicating

impacts of other confounders including level of education, health literacy, and cultural and

religious differences interfering with optimum cancer care for these Indigenous patients.

Breast cancer disparities in the United Kingdom

Compared with the USA, data from the UK on ethnic disparities in breast cancer are relatively

sparse (90-92). A recent study published with data from over 35,000 women with breast

cancer from South London reported that all minority ethnic groups in the UK have lower

breast cancer incidences compared with White European women (90). However, compared

with White European women, mortality hazard ratios were worse for women of all ethnic

minority groups, except Chinese. Adjusting for stage at diagnosis, treatment and

socioeconomic status almost completely explained these ethnic disparities in breast cancer

survival. Another study investigating breast cancer survival between South Asian and non-

South Asian women in the UK has reported South Asian women to have better 10-year

survival rates, which persisted after adjusting for age, stage and cancer treatment (92). Overall,

ethnic disparity in breast cancer outcomes in the UK appear to be much smaller compared with

both Australia and the USA, though they seem to follow a similar pattern.

Literature Review

33

3.2. Reasons for ethnic disparities in breast cancer outcomes

Several different explanations have been proposed to describe reasons for poor breast cancer

survival among Indigenous and ethnic minority populations. These explanations include more

advanced cancer stage at diagnosis due to delay in diagnosis and lower screening coverage,

more aggressive tumour biology, high rates of comorbidity and inferior quality of breast

cancer treatment. Most studies have analysed the impact of one or more of these factors

separately, and only a few studies have taken a holistic approach incorporating all these factors

included within contextual factors.

The following section explores the impacts of individual patient level factors, tumour factors

and healthcare service related factors on ethnic disparities in breast cancer mortality in New

Zealand and in other countries where similar disparities are encountered. Although I have used

this classification for ease of explanation, significant overlap among these categories for some

of the characteristics should be emphasized. For instance cancer stage at diagnosis will depend

on a combination of patient recognition of symptoms and presenting to a healthcare facility

(patient level), growth rate of the tumour (tumour level) and provision of an accessible and

equitable health service (healthcare service level) for patients (Figure 6).

3.2.1 Patient level factors

Patient level factors including cancer stage at presentation, comorbidity, socioeconomic status

and urban/ rural residency have been shown to have major impacts on cancer stage at

diagnosis and outcomes. As described above, some of these factors (i.e. stage at diagnosis)

will have an influence from tumour biology and healthcare service, while some other factors

(i.e. socioeconomic status, urban/rural residency) will influence accessibility of healthcare

services and treatment.

Stage at diagnosis

Stage at diagnosis (tumour size, nodal involvement and metastasis) is the strongest predictor

of breast cancer mortality (11). Compared to localized breast cancer with a 5-year survival of

over 90%, metastatic breast cancer remains largely incurable with a 5-year survival of

approximately 20% (93).


34

More advanced cancer stage at diagnosis in Māori compared with non-Māori is well known (4,

9, 69). All studies to date have shown that Māori were more likely to be diagnosed with more

advanced staged breast cancer compared with non-Māori women (Figure 7). This difference in

stage at diagnosis seems to have declined only marginally over last two decades (4). As

discussed previously, advanced stage at diagnosis is the biggest contributor to mortality

disparity, which is likely between a third and a half of the overall disparity. However, all

previous studies have used data from the NZCR, which includes a high proportion of unstaged

breast cancer (approximately 15%) complicating these estimations (15). A higher proportion

of Māori having an ‘unknown’ stage combined with possible inaccuracies in cancer staging in

the NZCR (94) may have resulted in an under estimation of the impact of advanced stage on

survival disparity between Māori and non-Māori women.

Figure 6: Model of pathways to cancer treatment (95)

HCP - Health Care Provider

The stage at diagnosis of a breast cancer is dependent on several of patient, tumour and health

care system factors (Figure 6), although the degree of impact of these factors for each woman

may vary substantially. This section discusses factors contributing to stage at presentation with

a greater emphasis on patient related factors, which include delay in presentation and

associated factors. Tumour and health system factors contributing to stage disparity are

discussed under respective topics.

Literature Review

35

Figure 7: Distribution of stage at diagnosis for breast cancer in New Zealand,1996-2001(Source –

Unequal impact – II) (69)

Barriers to access care

Differences in access to health care (defined as timely use of personal health services to

achieve the best possible health outcomes) is believed to be a major reason for ethnic and

socioeconomic inequities in breast cancer outcomes (96). Barriers to access may occur at

health system, health care process or patient level. Several international and local studies have

identified older age, minority ethnicity and socioeconomic status to be major barriers to access

optimum healthcare (96).

Disparities in access to healthcare for Māori compared to non-Māori have been demonstrated

across the New Zealand health care sector. Differences have been observed in terms of gaining

entry into services and differential experience of services including poorer quality of care and

a higher risk of adverse events and complications (35, 97-99).

Access to primary care has also been found to be poorer for Māori women with breast cancer.

Māori belief systems, such as views about reliance on the whānau (family), individual mana

(freedom or spiritual power), death and dying, and practices associated with tapu/noa (sacred

or holy), continue to influence health behaviour. These views may influence, for example,

preferences for care, individual help-seeking behaviour and responses to health care providers.

Previous qualitative research has identified cost of care, communication difficulties due to

factors such as low health literacy, structural barriers such as distance to travel and wait times,

previous negative experiences (e.g., institutional racism) and lack of ‘cultural fit’ in health

care services as barriers to access healthcare among Māori (100-102).


36

Level of education and health literacy are important factors associated with early diagnosis,

through promoting screening participation and minimizing delays in presenting to a primary

healthcare service for cancer related symptoms (Figure 8). Health literacy is defined as “the

degree to which individuals have the capacity to obtain, process and understand basic health

information and services needed to make appropriate health decisions” (103). The exchange

of health information is a complex process involving service configuration, the health

professional and the recipient (104). Key components of health literacy include individual

skills, health tasks undertaken, health materials used, skills of providers (including the ‘oral

exchange’), and the physical and social environment (105). Three out of four Māori females

have poor health literacy skills, which is approximately 50% higher than NZ European women

(106). It is likely that levels of education and health literacy are possible confounders

contributing to ethnic disparity in breast cancer outcomes not only through delay in diagnosis,

but also through interfering with optimum breast cancer treatment or through poor compliance

with treatment.

Figure 8: Cancer detection pathway (Source: The National Awareness and Early Diagnosis Initiative,

UK)(107)

Literature Review

37

Socioeconomic status

Several studies have shown that women of lower socioeconomic groups have lower survival

(Figure 9) and higher recurrence rates from breast cancer (108, 109). The higher breast cancer

mortality rate observed among patients of lower compared with higher socioeconomic groups

have essentially remained unchanged in New Zealand over the last two decades (6, 7). A

significantly greater proportion of Māori live in socioeconomically deprived circumstances

compared with NZ Europeans (Figure 10). Therefore it is believed that some of the observed

survival disparity between Māori and NZ European women is contributed by these

socioeconomic differences.

Figure 9: Age standardized mortality rates for breast cancer in New Zealand by deprivation quintiles

(Source – Ministry of Health)

Lower socioeconomic status has been shown to influence poor cancer outcomes through late

presentation, lower screening uptake, delays in treatment and poor compliance (110). Some of

the excess mortality observed among low socioeconomic patients could be explained by

differential access to healthcare services, although other associated factors including level of

education, health literacy, language and cultural differences also play a role (12). Similar

differences in breast cancer outcomes among different socioeconomic groups are observed in


38

countries with different health systems including the USA, the UK, Australia, Sweden and

New Zealand (111-115). Lower socioeconomic status which is more commonly seen among

Indigenous/ethnic minority women contributes, but does not fully explain the breast cancer

survival disparity (116).

Figure 10: Neighbourhood socioeconomic deprivation (NZDep2006) for Māori and non-Māori 2010

(Source: Tatau Kahukura, Māori Health Chart book 2010)

In the USA, lack of private insurance has been shown to be an independent predictor of

advanced stage at diagnosis and higher mortality, which has contributed to ethnic inequities in

breast cancer outcomes (117). Similarly, treatment provided in private sector which is

accessible to patients of higher socioeconomic status and/or with a private health insurance,

has also been shown to be associated with a shorter treatment delays and better long term

outcomes from breast cancer (118).

From a study conducted among women with breast cancer in New Zealand, McKenzie et al

demonstrated that the lowest socioeconomic group had a 50% higher breast cancer mortality

rate compared to the most affluent group, even after adjusting for ethnicity, tumour and socio-

demographic characteristics (12). Furthermore, the same authors in a later publication found

Literature Review

39

that the breast cancer survival disparity between Māori and non-Māori women was completely

attributable to social deprivation and differential access to health care, without any significant

contributions from cancer biology or treatment differences (13). However, other similar

studies from New Zealand suggest a much smaller impact from socioeconomic status, and

have demonstrated a significant residual ethnic disparity after adjustment for socioeconomic

status (7, 14). A study combining socio-demographic, tumour and treatment factors to clarify

and quantify the contribution of socioeconomic status towards the ethnic disparity has not

been done to date.

Urban / Rural residency

Residence in a rural compared to an urban area has been shown to negatively influence the

outcomes of cancer patients in different countries including the USA, Australia, France,

Canada and New Zealand (119-122). This disparity is mostly due to the advanced stage at

diagnosis in patients from rural areas, and this risk seems to increase proportionately to the

distance from residence to cancer centre, as reported by Campbell and colleagues from

Scotland (123).

International evidence on urban rural disparity in the stage at diagnosis and survival of breast

cancer is somewhat similar with most studies to date reporting rural patients to present with

advanced stage disease (124, 125). Rural women also experience poor breast cancer outcomes

even after adjusting for stage (126, 127). In contrary, a New Zealand study based on the

NZCR data for 1994-2004 period has reported that women residing in remote areas with

cancers of breast, colon and melanoma to be diagnosed with relatively early staged disease

compared with women residing in main urban areas (122). Despite that, this study has shown

significantly worse stage adjusted survival rates for rural women compared with urban

women, with a hazard ratio of 1.24 for most remote compared with urban women. Findings

from several other New Zealand studies have not supported those findings, as these studies

have failed to show significant differences in breast cancer stage at diagnosis or survival

between urban and rural dwelling women (15, 128, 129). Based on current evidence, it appears

that rural compared with urban residence is not associated with more advanced breast cancer

stage at presentation in New Zealand; however the impact of residential status on breast cancer

survival remains unclear.


40

Breast cancer screening

Mammographic screenings for early detection of breast cancer has shown to significantly

reduce mortality in women aged 50-69 years (130). Regular two-yearly breast screening has

shown to reduce the population risk of dying from a breast cancer by about 30% (131). A

recent review conducted by the Independent UK Panel on Breast Screening estimated the

relative mortality risk reduction to be 20% (95% CI 11%-27%) among women invited for

screening compared with women who were not invited (130). They further estimated that one

breast cancer death would be prevented for every 180 women screened (130).

Screening coverage of BSA has grown steadily since it was established in 1999. However, still

there are wide variations in the levels of coverage by different age groups, geographic areas

and ethnic populations. BSA achieved its target biennial coverage of 70% for NZ European

women in 2010 (132). However, for women aged 45 – 49 years it was 64.1% and for Māori

54.9%, both of which were well below the target 70% level (132). Further, there were large

variability in screening coverage rates for Māori by region, which ranged from 54% to 79%

across the country in 2012 (40). Lower screening coverage for Māori is a likely contributor for

more advanced disease, which in turn is the major contributor for lower breast cancer survival

in Māori (133).

Similar ethnic disparities in breast cancer screening participation rates have been reported

from the USA, the UK and Australia, with lower rates been observed in African Americans in

the USA and the UK, and Aborigines in Australia, compared with White women (134). Lower

breast cancer screening coverage has been shown to be responsible for approximately 10-25%

of the ethnic disparity in breast cancer mortality in the USA based on different statistical

models (83, 135).

Difficulty or barriers to access mammographic screening has shown to be a more important

factor for lower screening participation in Indigenous/ethnic minority women and in women

from low socioeconomic groups (136-139). In addition, uneasiness with the health care system

due to mistrust, perceived discrimination or perceived lack of quality have also shown to

impact on screening participation in ethnic minority women in minority African American and

Latina women in the USA (140, 141).

McNoe and colleagues studied reasons for non-participation in breast cancer screening during

the pilot phase of New Zealand national breast screening programme in 1996 (142). Practical

difficulties in attending screening including time, cost and transport, and negative attitude

Literature Review

41

towards screening among Māori including fear of the procedure or of a diagnosis of cancer

were some of the more common reasons for non-participation in Māori. Thomson et al,

published a retrospective process evaluation which evaluated the impact of several strategies

which were aimed at improving mammographic screening participation in a rural,

predominantly Māori population (143). Several community and practice-based strategies were

implemented with existing health resources through participation of Māori health providers

aimed at improving access, communication and community participation. These interventions

resulted in an increase in the rate of screening participation from 45% to 98% over a two year

period, which was observed to be maintained two years later. This study has demonstrated

some of reasons for screening non-participation in Māori and perhaps more importantly,

possibility of improving participation, with simple but well-planned strategic interventions.

Comorbidity

Comorbidity is the co-existence of diseases or disorders in addition to a primary disease of

interest. Comorbidity has many detrimental effects on cancer survival, the degree varies by

cancer site (144-146). These detrimental effects include lower use of primary and adjuvant

therapy, treatment delays and higher rate of treatment associated complications (147).

Breast cancer treatment is multimodal; for a majority of women it includes adjuvant systemic

therapy and/or radiotherapy in addition to surgery. Patient’s general health status and

comorbid illnesses may reduce tolerability and increase complications of primary and/or

adjuvant treatments, and may force to use a lower dose or a shorter duration of treatment or in

worst scenarios, a complete discontinuation of treatment.

Numerous studies have shown that comorbidities to be an independent predictor of worse

overall and cancer specific survival in women with breast cancer (148, 149). From a cohort

study of women with breast cancer with a 10-year follow up, Tammemagi and colleagues

showed that approximately 50% of the observed survival disparity among African Americans

and white women, to be due to differential distribution of comorbidities (146). However, other

similar studies from the USA have quoted much smaller contributions from comorbidities at

less than 25% (150, 151). Another US study by West and colleagues, failed to show any

significant impact of co-morbidities on survival disparity between African American and

White women despite the strong correlation observed between the comorbidity index score

(Charlson index) and mortality rate (152, 153).


42

A recently published New Zealand study has shown the impacts of different comorbidities on

cancer mortality including for breast cancer (154). This study which has used data from the

NZCR and the National Minimum Dataset (NMDS) has shown a direct relationship between

comorbidity and, poor overall and cancer specific survival. The impact of comorbidity on

mortality has been shown to be greater for cancers with a generally good prognoses including

for breast cancer compared with cancers with a relatively poor prognosis, for example for

cancers of liver and stomach (154).

Māori generally have a higher prevalence of comorbidities compared with non-Māori (33).

Comorbid illnesses such as cardiovascular diseases, chronic respiratory illnesses and diabetes

have been shown to be significantly higher among Māori, in comparison to non-Māori women

(68). This difference may be influenced by higher prevalence rates of modifiable risk factors

such as smoking, alcohol use and obesity in Māori compared with non-Māori women.

Comorbidity has been shown to be a significant factor for poor overall and cancer specific

survival for cancers including bowel and cervical cancer in Māori (67, 154-157). Although it

is clear that comorbidity is a likely contributor to ethnic disparity in breast cancer outcomes,

its proportional contribution is unclear at present.

Body mass index (BMI)

Obesity has a complicated relationship with risk of developing breast cancer and with clinical

behaviour of established disease. It is has been shown that obesity is associated with both an

increase in risk of developing breast cancer and a worse prognosis (158-160). It has also been

shown that the magnitude of difference for cancer recurrence in obese women versus lean

women to be comparable to the difference achieved by use of systemic hormonal and

chemotherapy (161).

Obesity is increasing at alarming rates in almost all countries in the world and has rapidly

become a major global health problem (162). It is believed that increasing obesity rates are

partly responsible for increasing post-menopausal breast cancer, especially in developing

countries. Obesity is one of the few modifiable risk factors that influence both development

and prognosis of breast cancer. It is suggested that reducing population prevalence of obesity

could reduce the number of cases of breast cancer by a tenth in Europe (163).

Rates of obesity have rapidly risen in New Zealand with one in every three adults being obese

currently (33), which is worse among socioeconomically deprived and ethnic minority

Literature Review

43

populations. In 2010, adult obesity rates were 48% and 68% respectively, for Māori and

Pacific compared with 30% for NZ Europeans (33). A recently published multi-ethnic case-

control study from New Zealand, has failed to demonstrate a differential impact of obesity on

breast cancer incidence disparity among Māori, Pacific and non-Māori non-Pacific women

(164). However, significant selection bias was documented for this study where response rates

for Māori and non-Māori women were 38.1% vs. 56.8%, respectively. This selection bias may

have impacted on the study results and its conclusions. No New Zealand study to date has

assessed the link between BMI and its impact of breast cancer mortality disparity in Māori

(and Pacific) women due to unavailability of BMI data from routine breast cancer datasets

including the NZCR or other regional breast cancer registries.

3.2.2 Tumour factors

Breast cancer is a heterogeneous disease; some are slow growing with an excellent prognosis,

while some have a grave prognosis irrespective of aggressive therapy. Measures of cancer

biology include;

Histological type

Tumour grade

Lympho-vascular invasion (LVI)

Oestrogen (ER) and Progesterone receptor (PR) status

HER-2 (Human Epidermal Growth Factor Receptor type 2) status

Histology, tumour grade and LVI

Differences in breast cancer biological characteristics by ethnicity are observed in many

countries, and between many ethnic groups. For example, African American women with

breast cancer in both the USA and the UK are known to have higher grade, ER/PR negative

and triple negative breast cancers than their European American counterparts (11, 165-167).

Further, these differences are known to be major contributors for excess breast cancer

mortality in African American women (11, 166). Some of the biological differences in breast

cancer between African American and White women have been attributed to lower

socioeconomic status, which is more common among African American women. Differences

in cancer biology between African American and White women appear to persist after


44

adjusting for age and socioeconomic status (168), which supports an independent association

between ethnicity and cancer biology.

Although several New Zealand studies have compared the ethnic differences among Māori

and non-Māori on tumour grade, no published study to date has compared the distribution of

histological varieties between the two ethnic groups.

A study carried out in Christchurch, New Zealand (total of 337 patients) showed a significant

difference in tumour grade between Māori and non-Māori, with more Māori women having

lower grade tumours (169). These findings contradicted a study published by the Auckland

Breast Cancer Study Group with a larger sample of 1577 women, where they found Māori

women to have significantly more, higher grade tumours (74). Further, they postulated that

this difference in the grade being a contributor to the higher mortality from breast cancer

observed among Māori women. These findings were further supported by McKenzie et al who

published data from 2968 patients, where Māori women had significantly more, higher grade

tumours compared to non-Māori (13). However, tumour grade was not observed to have a

significant effect on the mortality disparity between Māori and non-Māori, when adjusted for

other tumour variables including tumour size and extent of spread.

Very few studies to date have compared the impact of LVI in breast cancer outcome among

different races/ethnicities in the world. Chavez-MacGregor and colleagues compared the rates

of LVI among ethnicities during their study on breast cancer response to neo-adjuvant

chemotherapy (170). Their data showed that there was no significant difference in rates of LVI

among African Americans, Hispanics and White women. Weston study based the Auckland

Breast cancer Register, reported that there was no difference in the rates of LVI among Māori,

Pacific and non-Māori, non-Pacific populations (74).

ER/PR and HER-2 status

Women with ER and PR negative tumours have approximately a 10-15% lower 5-year disease

free survival rate than women with ER/PR positive cancers (171-173), while HER-2 amplified

tumours (not treated with trastuzumab) have been shown to be associated with approximately

a 5-10% lower 5-year survival compared with HER-2 non-amplified cancers (172). A US

study using Surveillance, Epidemiology and End Results (SEER) data from over 70,000

patients’ demonstrated significant differences in tumour histology and receptor status by

ethnicity which were also observed to be linked strongly with breast cancer mortality (86).

Literature Review

45

These findings were supported by several other studies including the Women’s Health

Initiative (WHI) study and Carolina Breast Cancer Study from the USA, which demonstrated

significantly higher proportions of ER/PR negative (including triple negative breast cancer) in

African American women (11, 165, 166). Further, a recent publication has shown a 50%

higher incidence of triple negative cancer in African American compared with White

American women, while no differences in HER-2 status were observed among different ethnic

groups (174).

ER/PR status has been measured routinely in New Zealand since early 1990’s, but no major

differences in ER/PR status has been observed among ethnic groups (13, 74, 169). However,

some of these studies were too small to detect a statistically significant difference, if one

existed and others based on the NZCR might have been influenced by a high proportion of

missing data.

HER-2 receptor status has only been routine in breast cancer pathology reporting in New

Zealand over the last 7-8 years. Analysis of data from the NZCR has shown a higher rate of

HER-2 positive cancers among Māori and Pacific compared to non-Māori, non-Pacific women

(13). However, HER-2 status has not been shown to be associated with breast cancer mortality

disparity (13). A high proportion of missing data (38%) in this study has weakened the

accuracy of these conclusions.

Overall, present evidence suggests significant differences in breast cancer biological

characteristics by ethnicity in New Zealand. These appear to be dissimilar to biological

differences reported from the USA. It is unclear at present, whether these biological

differences are a factor contributing to mortality disparity, similar to the contribution seen for

African American women in the USA.

3.2.3 Healthcare service factors

Relatively few published papers have directly analysed ethnic disparities in healthcare access

or utilization in New Zealand. The studies that have been published have reported significant

disparities in care between Māori and non-Māori patients for diabetes (102, 175), cardiac

interventions (176-178), surgery (179, 180) and for cancer treatment (181, 182).

In 2002, the National Diabetes Working Group conducted a literature review to investigate

barriers to access care for Māori with diabetes (102). This review identified cost of care, low


46

numbers of Māori among health staff, and lack of understanding among mainstream health

workers concerning the context in which Māori patients live their lives as major barriers for

optimum care for Māori. A study from South Auckland included in the previously mentioned

review, has reported that lack of community based services and lack of perceived benefit from

treatment as common factors for poor compliance among Māori and Pacific patients (175).

Several studies have investigated and reported on the lower rate of cardiac interventions in

Māori who have a much higher rate of ischaemic heart disease and cardiac deaths compared

with non-Māori (176, 177). For instance, during the 10-year period between 1990 and 1999,

rate ratios for coronary artery bypass grafting and percutaneous coronary angioplasty for

Māori compared with non-Māori non-Pacific were 0.40 and 0.29, respectively (183).

However, these disparities seem to have improved substantially, especially over last 5-10

years as reported by Kerr and colleagues (178). This study has not shown a significant

difference in rates of cardiac intervention by ethnicity during 2007-2012, once adjusted for

other disease characteristics.

Reported ethnic disparities in cancer care in New Zealand include care for colon and lung

cancers. Stevens and colleagues have reported on fewer curative resections and longer

treatment delays for surgery for lung cancer experienced by Māori compared with non-Māori

patients in Auckland and Northland regions (182). A more detailed analysis of disparities in

colon cancer care in New Zealand has been done by Hill and colleagues. They conducted a

nationwide cohort study of colon cancer disparities using a case note review, and reported that

Māori patients to have received substantially poorer quality of cancer care compared with non-

Māori patients (67, 181, 184). These included lower likelihood of undergoing a curative

resection for operable cancer, higher likelihood of undergoing emergency surgery and higher

likelihood of undergoing a palliative resection compared with non-Māori. Further, eligible

Māori were less likely to be offered chemotherapy and were more likely to experience longer

delays to receive chemotherapy compared with non-Māori patients. This study concluded that

disparities in cancer care have been responsible for approximately a third of colon cancer

disparity between Māori and non-Māori patients.

These findings are further supported by data mostly from the USA on ethnic disparities in

cancer treatment and its impact on outcome disparities. Several researchers in the USA have

studied ethnic differences in surgical treatment of cancer, which included surgery for cancers

of lung, oesophagus, prostate, liver, endometrium and colon (185-190). These studies have

reported statistically significant differences in rates of surgical treatment for respective cancers

Literature Review

47

between African American and White American patients. Perhaps more importantly, these

studies have shown minimal or no survival difference between patients from these two ethnic

groups with similar staged disease who underwent surgery.

To assess the ethnic disparities in breast cancer treatment in the USA, Li and colleagues

carried out a study with 124,934 patient data from SEER database covering 11 regional breast

cancer registries over a 7-year period (80). They found that patients with breast cancer from

ethnic minorities were more likely to receive surgical or radiation therapy not meeting national

standards. The authors have concluded that these differences were more likely to be due to

socioeconomic and cultural backgrounds of these patients, rather than due to discrimination by

health care providers. However, whether these differences in care represent differences in

patient preference, provider decisions, poor patient-provider communication or a deficiency of

the healthcare structure has not been studied adequately in context of cancer surgery.

Shavers and Brown published a review of literature on ethnic disparities in cancer treatment in

the USA (191). Their review has reported significant ethnic differences in cancer treatment,

and has identified several structural barriers as well as physician and patient related factors

contributing to ethnic differences in treatment. General observations following their study of

23 publications on ethnic disparities in breast cancer treatment in the USA were that, ethnic

minority women to have a lesser chance of receiving breast conservation surgery (BCS) and

radiotherapy following BCS, a lesser chance of receiving adjuvant chemotherapy or hormonal

therapy and a higher likelihood of receiving chemotherapy and hormonal therapy for a shorter

duration not keeping in with recommended guidelines compared with Caucasian women (192,

193). However, all the studies were not in agreement and there were several studies that

contradict above general observations, which have demonstrated absence of any significant

differences among different ethnicities in rates of BCS, radiotherapy, chemotherapy and

hormonal therapy (194, 195). Differences in study methodology, smaller sample sizes and

possible regional variations in cancer care might have contributed to these variations.


48

Quality of treatment

Early detection through screening mammography or through general practitioners and timely,

optimal treatment provided through a multi-disciplinary team (MDT) is believed to be the

ideal method for achieving best long term outcomes from breast cancer (196).

Major advances in treatment of cancer have been made over past few decades. These advances

have resulted in significant survival improvements for most cancers. Breast cancer survival

also has shown a marked improvement, mainly for localized cancer which is over 90% at five

years in the developed world (93). However, there are clear differences in survival

improvement among different ethnicities in many countries, indicating unequal distribution of

advances in cancer therapy. Many a research has looked into the impact of possible ethnic

disparities in the administration of cancer treatment and provider delays as contributors to this

disparity.

Possible areas of lower access or inferior quality of care for Māori compared with non-Māori

include; delays in diagnosis and treatment (including primary and adjuvant therapy), inferior

quality or inadequate use of appropriate treatment (due to lower referral rates, comorbidities,

patient declining treatment) and a lesser proportion of patients being managed through MDTs.

In New Zealand, BreastScreen Aotearoa (BSA) Independent Monitoring Reports and Māori

Independent Monitoring Reports have published data on treatment differences for screen

detected breast cancer by ethnicity, and to date are the only reports to have published such data

(40, 197). BSA Independent Māori Monitoring Reports were commissioned to provide an

analysis of Māori data, and inequities between Māori and non-Māori for quality and timeliness

of assessment and treatment for women diagnosed with screen detected breast cancer (197).

These reports have documented significant inequities in breast cancer treatment between

Māori and non-Māori women. For example, Māori women were 20% less likely to receive

their first treatment surgery within 20 working days than non-Māori women.

Overall inequities along the screen detected breast cancer treatment pathway between Māori

and non-Māori women appear to be minimal, especially in comparison with what is known of

non-screen cancer treatment pathway (181). It is possible that inequities have been minimised

due to intensive quality control measures along the screen detected breast cancer treatment

pathway, including monitoring against key quality indicators, audit and independent review.

For non-screen detected cancers the disparities are likely to be greater, but we lack this

information, which is a major deficiency.

Literature Review

49

Delay

A delay in presentation with symptoms of breast cancer is known to reduce the rate of survival

from breast cancer (198). A total delay, which is a combination of patient and provider delay

(Figure 11) of more than three months from the time of detection of symptoms to initiation of

treatment has been shown to be associated with significantly higher rates of mortality and

cancer recurrences (199). In the only published New Zealand study to date looking at delay in

relation to breast cancer, Meechan and colleagues reported that 13% women presented with a

delay of longer than three months. However, the authors have not assessed the differences in

delay among different ethnicities in their study population (200).

Figure 11: An illustration of the overall milestones and time intervals in the route from first symptom

until start of treatment (201)

Several international researchers have studied ethnic differences in delay in breast cancer,

mainly focusing on patient delay. Facione and colleagues performed a critical review of

literature on ethnicity and delay in breast cancer presentation in the USA using 12 published

studies (202). They concluded that there is a consistent and significant longer delay among

African Americans and Hispanics compared with White women across all studies. However,

possible reasons behind these delays were less clear, but possibly included lower level of

education, poor socioeconomic status and lack of ready access to health care. Other

researchers have investigated provider delay in cancer therapy, among different ethnicities in


50

the USA. These studies have shown that there are longer provider delays in treatment for

African Americans and Hispanic breast cancers, compared with White Caucasian women

(203, 204).

In a meta-analysis published in the Lancet in 1999, Richards and colleagues confirmed the

relationship between delay in treatment and poor breast cancer survival (198, 205). In this

meta-analysis, a delay of more than three months from the onset of symptoms to initiation of

treatment was associated with a 12% lower 5-year survival compared with a delay of less than

three months, after adjusting for lead time bias. A recent study published by McLaughlin and

colleagues have shown that a much shorter delay of 2-months was associated with

significantly higher mortality rates for advanced breast cancers and for cancers associated with

adverse prognostic characteristics including higher grade, HER-2 positivity and triple negative

cancers (206).

According to BSA quality standards, at least 90% women should receive their first surgical

treatment within 20 working days of receiving their final diagnostic result (132). However,

figures from BSA in 2008 indicate that only 57.7% Māori and 71.2% non-Māori women

achieved these targets. Given the lack of standards and audit for management of symptomatic

non-screen detected cancers, disparities are likely to be greater, although we lack data on this

currently.

Timeliness of instituting adjuvant treatment has also been shown to be crucial for the

maximum potential benefit from these treatments. Two recent meta-analyses have shown a 6%

and 15% increase in relative mortality rate with each 4-week delay in initiating adjuvant

chemotherapy for breast cancer (207, 208). Delays longer than 3 months for adjuvant radiation

therapy has been shown to be associated with higher risks of local recurrence and mortality

(209). Although timeline thresholds given in treatment guidelines are sometimes arbitrary and

controversial, longer delays for surgery, chemotherapy and radiation therapy have all been

proven to be associated with poorer breast cancer outcomes including higher risks of

recurrence and mortality (198, 207-210).

Several New Zealand researchers have looked into possible provider delays, leading to worse

cancer outcomes in Māori. In two such studies on treatment differences for lung and colon

cancers, authors have concluded that Māori were more likely to experience significantly

longer delays for treatment compared to non-Māori patients (181, 182). They have also

highlighted the higher frequency of defaulted appointments and delays in consent for

treatment by Māori, contributing to observed provider delay. However in a publication in the

Literature Review

51

Lancet in 2006, Harris et al claimed that ethnic discrimination in New Zealand public health

institutions, as a major reason for such defaulted appointments and delays in consent by

Māori, which adds up towards the ethnic disparity in outcomes for major diseases including

cancer (35). This is further supported by two previous studies on asthma and diabetes which

have shown that either conscious or unconscious attitudes of healthcare workers have

contributed to delay in seeking medical care among Māori (175, 211).

In an effort to reduce delays in cancer treatment the Ministry of Health introduced Faster

Cancer Treatment Indicators as targets for timeliness of cancer treatment for all District Health

Boards in 2012 (58). According to these indicators cancer treatment is expected to be initiated

within 62 days of initial referral and within 31 days from the day a decision to treat the cancer

is discussed with the patient. Whether this initiative could reduce overall delays in cancer

treatment and ethnic disparities in delay is yet to be seen.

Adherence with treatment and other patient related factors influencing treatment

Completion of cancer treatment including adjuvant therapy and a proper follow up after initial

treatment plays an important role in improving overall breast cancer outcome. Evidence from

population based studies indicates that poor compliance with adjuvant therapy and missed

follow up appointments are important factors negatively influencing breast cancer survival

(212). The reasons for differences in compliance among ethnic groups are complicated and it

depends on a multiplicity of patient, physician and system related factors (213, 214).

Indigenous populations from different parts of the world have varying views on diseases

including cancer that sometimes significantly differ from the western beliefs. Similarly, Māori

also have a different view on diseases and treatment, which is based on their culture and

beliefs (214). When a Māori woman with breast cancer has to obtain allopathic treatment from

a non-Māori provider it has the potential for a conflict with their beliefs and culture, which

could result in poor compliance with prescribed treatment. Several studies from New Zealand

support this theory, as they have shown poor compliance among Māori with treatment for

several chronic illnesses including cancer (157, 215).

One example of importance is adherence for adjuvant endocrine therapy for hormone receptor

positive breast cancer. Endocrine therapy has been shown to reduce risks of cancer recurrence

for early breast cancer by approximately a half and breast cancer mortality by approximately a

third (216). However many women do not get the maximum potential benefit either due to


52

non-prescription or due to discontinuation of treatment prematurely or due to not maintaining

an optimum level of adherence over the duration of treatment, which result in higher

recurrences and lower survival rates (217).

Data from the National Breast Cancer Audit of the Royal Australasian College of Surgeons

showed that only 81% of NZ women with oestrogen receptor (ER) positive invasive breast

cancer to have been prescribed with endocrine therapy in 2008 (218).

Non-adherence with endocrine therapy is common (25-35%) and, is one of the most important

areas of sub-optimal breast cancer care (219). Side effects of endocrine therapy, lack of

understanding of its benefits which is related to poor health literacy, comorbidities and cost are

known barriers to adherence (220). Lower adherence to endocrine therapy has been observed

more often among women from ethnic minority groups and lower socio-economic groups in

the USA (221). Further, improving adherence to endocrine therapy is an area where substantial

cost effective gains are achievable. Several studies have proven the effectiveness of measures

to improve adherence with endocrine therapy such as increased patient awareness of its

benefits, regular reinforcement of advice and providing supportive care to manage side effects

(222).

Multi-disciplinary Team care (MDT care) and Cancer Care Coordinators (CCC’s)

Providing breast cancer treatment through a dedicated multi-disciplinary breast cancer care

team has been shown to increase the overall quality of care and survival from breast cancer

(196, 223). In keeping with this, the Ministry of Health has implemented several initiatives

targeted at increasing the proportion of cancers provided with care through MDTs including

for colo-rectal, breast and lung cancer.

Recent introduction of identification and flagging of patients at risk for poorer outcome (e.g.

due to low socioeconomic status, ethnicity, poor family/social support) through MDTs has met

with some success for colo-rectal cancer patients. This strategy is aimed at proactive

identification of at risk patients during early part of cancer management and providing them

with support to achieve optimal outcomes. However, this requires a comprehensive knowledge

of factors associated with poorer outcomes especially those factors which can be most easily

improved by healthcare service interventions.

Literature Review

53

Improved patient navigation through dedicated cancer care coordinators (CCC) has been

shown to help reduce delays, especially for women who are at-risk for longer delays which

include women of minority ethnicity and low socioeconomic groups (224). These nurses act as

a single point of contact for cancer patients. CCCs help to coordinate care, providing

continuity of care and support from diagnosis through the course of cancer care. Personalised

coordinated care programmes for cancer patients have been shown to improve timeliness of

care (225, 226), and patient satisfaction with the level of care and support (227). CCC’s have

also been shown to reduce inequity in access to care by reducing barriers relating to cultural,

language, educational, socioeconomic and geographical factors (228) and thereby to improve

quality of the delivery of nursing care.

The Waikato DHB has two full-time CCC’s providing support for women with breast cancer

since 2009. The Ministry of Health has identified CCC’s as a key strategy to increase quality

and reduce inequalities in cancer care, and since 2012 has provided funding for all District

Health Boards to employ CCC’s for management of common cancers in New Zealand (229).

Institutional discrimination

Discrimination based on patient ethnicity, socioeconomic status or residence could impact

significantly on these patients who are anyway at risk of worse cancer outcomes. Following a

population based study in New Zealand, Harris and colleagues reported that previous

experiences of ethnic discrimination in public health setup as an important factor leading to

lack of participation of Māori women in cervical and breast screening programmes (230).

Despite the absence of further studies to support their claims, the authors have raised an

important issue that many have ignored in investigating reasons for ethnic disparities in breast

cancer outcome in New Zealand. This issue and its possible impacts are discussed further in

next section.


54

3.3. Understanding key drivers behind ethnic inequities in breast cancer

outcomes

Ethnic inequities in breast cancer outcomes in New Zealand appear to be a direct or indirect

result of differences in healthcare access and quality of treatment received by Māori compared

with NZ European women. This raises the issue as to why Māori women experience greater

barriers to access healthcare and once gained access, a quality of care which possibly is

inferior to NZ European women. New Zealand has a publicly funded health care system that

provides free hospital and specialist care for all citizens with the aim of providing equitable

healthcare for all patients irrespective of ethnicity or income. Yet, Māori appear to experience

a poorer quality health service compared with NZ Europeans.

Inequities in cancer care are generated within the healthcare system both due to its structural

organization as well as due to the way in which healthcare services are delivered. For instance

geographical location or segregation of healthcare services could result in longer travel

distances or preferential access to some patient groups over others. Further, health insurance

status, type of the hospital where care is received and regional variations in quality of care also

could contribute to inequities in treatment as a consequence of structural organization of health

services (191). Treatment recommendations for a woman diagnosed with a breast cancer

would be based upon cancer stage, biological characteristics, comorbidities and patient

preference. However, physician perceptions and physician bias also could influence these

decisions. Further, patient perceptions, beliefs and expectations, and social and socioeconomic

circumstances may influence the treatment process directly or through interactions with

physicians and healthcare institutions (191).

Several authors have attempted to disentangle underlying causes for ethnic inequities in cancer

care and have proposed different frameworks and theories. Of these, two reviews by Shavers

et al and Mandelblatt et al and a report published by Smedley et al standout (96, 191, 231).

These two reviews have used similar frameworks to categorise factors into healthcare system

factors, factors influencing physician decision process and patient factors, while the report by

Smedley et al has categorized factors into system, provider and patient factors (Table 2).

Literature Review

55

Table 2: Potential barriers interfering with optimal cancer treatment

Category Specific factors

Healthcare structure (System level) Geographical location and type of institution

Healthcare funding

Physician level factors (Provider level) Physician perceptions / Biases

Communication and cultural safety

Patient level factors Clinical factors

Patient preference / Choice

Social and socioeconomic circumstances

The following section discusses and explores mechanisms contributing to ethnic inequities in

access and quality of treatment under broader topics of healthcare structure, physician and

patient related factors while acknowledging their significant overlaps and mutual interactions.

3.3.1 Healthcare structure (System level)

The structure of a healthcare organization should be designed in a way to maximally facilitate

the provision of care needed for patients, including for cancer. However, organizations of

many healthcare systems induce potential barriers to optimum level of care, and include

organizational and structural factors, funding and financial forces and regional factors.

Shavers et al, in their review of ethnic disparities in cancer care have identified health

insurance as a major system level determinant of ethnic disparities in receipt of healthcare in

the US setup, which lacks a proper publicly funded healthcare system (191). A greater

proportion of minority Black Africans compared with White in the USA, either lack health

insurance or do not have an adequate health insurance policy, which significantly hinders

access and to receive quality healthcare. In a fee levying health system, as in the USA, these

healthcare access disparities are greater for conditions such as cancer, where the costs are

generally greater (232). Other system level factors identified in the USA include provider

resourcing, cultural focus of healthcare services and segregation of healthcare services (231).


56

Access to Cancer Services for Māori, a report compiled by the Ministry of Health has

documented several healthcare system level factors creating difficult access and inferior

quality of cancer care for Māori (97). This report has identified four main areas contributing to

these disparities and includes location of cancer services, cost of cancer care, focus of cancer

services, and the composition of the cancer service workforce. These findings are similar to

the findings reported by Baxter and colleagues for diabetes care in New Zealand where

provision of a health service that does not cater to cultural and societal needs of patients was

found to be a major factor (102).

Geographical location

Healthcare service institutions are generally located or concentrated in urban areas which have

a high population density. New Zealand in general and to a greater degree in the Waikato,

includes large proportions of rural and remote areas. For instance, the Waikato DHB covers an

area of over 20,000 square kilometres, yet has only a single major city, Hamilton, where the

tertiary hospital providing specialist cancer services is located. This means that some remote

dwelling patients in the Waikato have to travel up to 200 kilometres to access specialist cancer

care. Differential urban rural distribution of Māori and NZ European populations in New

Zealand and in the Waikato means that this invariably contributes to healthcare access inequity

between Māori and NZ European patients.

In 1990s, many changes to the health care structure were instituted in New Zealand, and

included downgrading or closure of many rural hospitals and centralization of specialist cancer

services to larger metropolitan hospitals (233). Although the impact of these changes on

access and quality of healthcare for rural and remote populations have not been studied, they

obviously have increased travel times and costs, and perhaps longer wait times to access

centralized cancer care.

Many healthcare practitioners have identified geographical location or service location as

significant issues influencing healthcare service access for Māori as reported in Access for

Cancer Services for Māori (97). Lack of or inadequate locally accessible healthcare services

probably have affected Māori more than NZ Europeans as this has created a barrier for

whānau-based care, which is the expectation among many traditional Māori patients.

Healthcare provision based on ethnic segregation of communities is seen prominently in the

USA, and is believed to be a factor contributing to ethnic inequalities in cancer care. Although

Literature Review

57

official segregation of care does not exist anymore, in reality, clear differences in the quality

of care provided to Black compared with White communities are observed in the USA (234).

Such differences in care are driven mostly by differences in healthcare insurance, which has

resulted in better hospitals being located within rich White communities who have better

health insurance, while low quality hospitals are being located within Black communities, who

are covered mostly by government provided medical insurance.

However, little is known about the impact of health insurance on quality of healthcare or

whether there is a difference in quality of service provided by private hospitals which provide

care for patients with health insurance, compared with the public system. One such study

comes from neighbouring Australia which has reported on relatively poorer quality of

healthcare received by public compared with private patients for lung and breast cancer, which

was associated with poor long term cancer outcomes (118). Despite the lack of evidence from

New Zealand, presence of disparities based on private public private care in Australia, which

has a healthcare service similar to New Zealand, strongly indicate the existence of similar

disparities in New Zealand. As fewer Māori are known to possess a health insurance policy

compared with NZ Europeans (235), this may have created another avenue through which

disparities in care are induced, contributing to cancer outcome disparities between Māori and

NZ Europeans.

Provision of healthcare through greater Māori provider participation for predominantly Māori

communities is considered as a possible way of improving accessibility and providing a better

quality service for Māori patients. A similar approach has been reported to be highly

successful in improving mammographic screening participation for Māori (143). However,

provision of healthcare through Māori providers represent only a small proportion of

healthcare delivery presently which is limited almost exclusively to primary health care

services (36).

Healthcare funding

Hospital and healthcare expenditure have sky-rocketed especially over the last two decades.

Reasons for this increase include increased man-power costs, use of advanced technological

equipment, and use of newer more costly treatment. Expanding population, especially the

elderly population, with added healthcare needs have further exacerbated the healthcare costs.

In the USA, this has resulted in closure or restructuring of many hospitals to maintain financial


58

viability (236). In New Zealand, as a greater portion of healthcare delivery rests with the

public system which is government funded, the implications of cost have been different from

the USA. Healthcare reforms in 1990s were introduced to absorb some of these financial

strains by reducing cost of healthcare, but have resulted in some major and significant quality

issues. For instance, major shortages in provision and delays in radiotherapy for cancer was

observed in early 2000s where over 40% of patients experienced delays longer than

recommended guidelines (237). Although wait times for cancer care seems to have improved

over the last decade, some of the centralized cancer treatment units still experience significant

staff and equipment shortages resulting in longer wait times for cancer care.

The primary health care system in New Zealand is highly subsidized, but patient co-payment

is also substantial. A visit to a general practitioner may cost between $20 and $50 for an adult,

and for cancer patients requiring repeated visits this may create a major financial constrain.

Cost of primary care has been shown to be the major reason for not visiting a general

practitioner when required and has been shown to affect Māori more than NZ European

patients (34). While some of this disparity is due to higher proportions of Māori within lower

socioeconomic categories, lower rates are seen for Māori even within the same socioeconomic

decile, indicating the presence of further barriers to primary care for Māori. Costs of primary

care is likely to further exacerbate financial constraints of cancer patients, due to additional

costs of cancer care including travel for cancer treatment, at a time when income may be lost

due to being away from work.

3.3.2 Physician level factors (Provider level)

Healthcare service expectations of Māori may significantly differ from beliefs and attitudes of

a non-Māori physicians providing care in a European/Western designed healthcare institution.

Poor communication due to cultural differences, knowledge, attitudes and behaviours may

further exacerbate this discordance, ultimately resulting in a quality of care which falls well

below patient expectations and perhaps, recommended standards. These factors may create an

impression in Māori that they are being subjected to discrimination, and rates of self-reported

discrimination in Māori which is about ten times higher than NZ European patients, support

this premise (238). Perceived discrimination among Māori patients is also likely to be

transferred to other members within the community creating a general reluctance to seek care

from mainstream public healthcare institutions.

Literature Review

59

Physician belief and attitudes

Health care providers play a pivotal role in ensuring access to and quality of cancer care for

their patient populations, and provider recommendations are one of the most consistent

predictors of receipt of early detection and other cancer services (239). Findings from several

studies from the USA suggest that a physician’s perception of patients may be influenced by

non-clinical characteristics including ethnicity, religion and socioeconomic status (240, 241),

which may then be manifested in differences in patient referral patterns and treatment

recommendations (241, 242). In a survey of physicians from the USA, it was reported that

physicians were more likely to have negative perceptions of African Americans and persons of

low or middle socioeconomic status than of Whites and persons of high socioeconomic status,

respectively (240). Differences in attitudes of physicians towards patients may impact on

patient care, and these perceptions are likely to contribute to ethnic disparities in cancer

treatment and resulting differences in cancer outcomes.

Physician beliefs could influence his/her decision making process through two main

mechanisms. First, the physician may believe that a particular group of patients (ethnic or

socioeconomic) are non-compliant, late-presenting and defaulting clinic visit type of patients

due to stereotyping (243). Barriers to access care including cost, travel, time off work and

caring for dependants would explain most of such behaviours, but unfortunately creates a

prejudice among some physicians due to lack of understanding of these circumstances. On the

other hand some other physicians may make decisions for these patients genuinely believing

that they are making the best decision for that particular patient group, but due to lack of

understanding of their values and culture, makes a decision that is either not optimal or not

acceptable to the patient. Whatever the mechanism that is involved it finally ends up creating a

process that leads to sub-optimal quality of care for these patients.

Communication and cultural safety

Physician-patient communication is another key domain in determining access to care. The

quality of communication on cancer care has been noted to vary by physician gender, by

patient race or ethnicity, and by patient’s socioeconomic status (244-246). For example, in the

USA physicians discuss mammography less often with their Hispanic patients than with their

non-Hispanic patients, and Black patients are less likely to receive advice about cancer

screening in general than Whites who see the same physician (247).


60

In New Zealand, similar quality issues in relation to communication between Māori patients

and health care providers have been reported. For instance, at primary care level, general

practitioners on average have shorter consultations with Māori and have lower levels of

rapport with Māori compared with NZ European patients (99). This contradicts with

expectations of Māori patients which include desire for careful listening, personal engagement

and face-to-face communication with health professionals (248). Further, Māori patients have

an expectation for them to be identified as Māori which enable them to participate in health

decision making and service engagement while maintaining their cultural and ethnic identity

(249).

Ensuring cultural safety provides the essential foundation that creates a suitable background

for effective communication. A commonly accepted definition of cultural safety is ‘an

environment which is safe for people; where there is no assault, challenge or denial of their

identity, of who they are and what they need’. It is about shared respect, shared meaning,

shared knowledge and experience, of learning together with dignity, and truly listening (250).

For Indigenous people including Māori, cultural safety is essentially a basic right recognised at

international and local levels (251). Lack of recognition and emphasis placed on this important

issue, however, appeared to have interfered with effective communication with Māori patients

within the mainstream healthcare system.

3.3.3 Patient level factors

Clinical factors

Patient clinical factors including comorbidity, obesity and smoking are known to influence

cancer treatment decisions mostly due to reduced tolerability of treatment or due to risk of

significant complications. Indigenous and ethnic minority patients are more likely to have

higher levels of comorbidity, obesity and smoking and hence are at a greater risk of not

receiving optimum cancer treatment (31). Studies have shown that Indigenous and ethnic

minority patients including Māori, with these risk factors to have received sub-optimal care

which is not adequately explained by presence of these conditions (67). Comorbidity and

obesity in many occasions are not contraindications for primary or adjuvant breast cancer

treatment, although change in regimens may be required (252). It appears that presence of

these risk factors has influenced treatment decision process, with a greater preference for not

to use optimum therapy, especially for Indigenous and ethnic minority patients.

Literature Review

61

Patient choice

Health beliefs and attitudes influence health-care behaviour (253). On the other hand, health

beliefs and attitudes are influenced by culture, education and health literacy of patients. For

example, the prevalence of fatalistic and nihilistic attitudes towards cancer have been shown to

be higher among ethnic minority populations than White populations in the USA (254). A

proportion of the higher treatment refusal rates reported among ethnic minorities in the USA

has been demonstrated to be due to differences attitudes towards treatment among these

populations (255). Treatment refusal however, may also be influenced by number of other

factors including poor provider-patient communication, lack of trust, provider patient

differences in beliefs and expectations and due to competing priorities.

Several New Zealand studies have reported on higher rates of cancer treatment refusal in

Māori compared with non-Māori patients (4, 182). However, as noted in the Unequal Impact –

II report, patient choice is only a minor contributor for treatment differences and is an unlikely

cause for cancer survival disparities in New Zealand (4).

Social and socioeconomic circumstances

Lower socioeconomic has a direct correlation with poorer cancer survival for many cancers

including breast cancer (191). Poor cancer outcomes among Māori is known to be influenced

to an extent by the higher prevalence of lower socioeconomic status among Māori compared

with NZ Europeans.

Lower social or socioeconomic circumstances are known to interfere with proper access to

healthcare services and are believed to be a major reason for poorer cancer outcomes in these

patients (4, 34). In the USA, socioeconomic status is closely correlated with health insurance

which determines access and choice of healthcare services (84, 118). Lack of time off work,

travel time, cost or lack of transport and lack of social support to care for dependants are some

of the common reasons influencing patients of lower socioeconomic states not to seek medical

care or delay seeking medical care. Although specialist care in public sector is provided free of

charge in New Zealand, co-payments involved with primary care and payments required for

specialist care in private are some of the other reasons, which may influence Māori patients to

receive an inferior quality of care compared with NZ European patients (31).


62

3.4. Conceptual framework

The discussion presented up to this point suggests that a considerable number of potential

causal explanations for inequities in breast cancer outcomes between Māori and NZ

European women. Drawing from the literature, a conceptual framework for Māori - NZ

European breast cancer inequities was developed and is presented in Figure 12. It is evident

from the literature that ethnicity and socioeconomic status are intermingled in their effect on

stage at diagnosis and survival/mortality; as such the conceptual framework illustrates the

pathways in which these two variables influence stage and survival outcomes.

Ethnicity and socioeconomic status could be considered to be influencing stage at diagnosis

and survival/mortality through three main mechanisms; patient characteristics, cancer biology

and the healthcare system/treatment.

This thesis aims to bring together all these characteristics in an attempt to provide explanations

for observed breast cancer inequity between Māori and NZ European women. Most of these

factors could be influenced by access to the social determinants of health, including racism,

and access, timeliness and quality of health care which can be influenced by health services

inequities. Initially, the impacts of ethnicity and socioeconomic status on cancer stage at

diagnosis, screening participation, cancer biological characteristics and treatment differences

are explored in detail to identify inequities in these key areas. The final analysis attempts to

bring together all findings from previous analyses into a single model to describe and quantify

the impacts of each of these factors towards the observed survival inequity between Māori and

NZ European women.

Literature Review

63

Patient factors

Acceptance

Health system Factors

Provision

Biological factors

Figure 12: Conceptual framework depicting complex interaction of patient, health system and cancer related factors influencing clinical outcomes

Ethnicity Socioeconomic

status

Screening

participation

Screening

programme

eligibility

Cancer biology

Comorbidities

Symptoms, presentation

& diagnostic process

Investigations

performed

Cancer stage at

diagnosis

Treatment offered

- guidelines

Treatment accepted

and completed

Supportive care

& follow up

offered

Clinical outcomes

from breast cancer

Supportive care

& follow up

accepted

Use of general

practice

Accessibility

of general

practice


64

Design and Methods

65

Chapter 4. Design and Methods

4.1. Study population

The target population for this study comprised of all newly diagnosed women with primary

incidental breast cancer while being a resident within the Waikato District Health Board

(DHB) area during a 14-year period from 01/01/1999 to 31/12/2012. Date of diagnosis was

defined as the first date of obtaining tissue/cells from the primary tumour or its secondary

deposits for confirmation of diagnosis of the index breast cancer.

Target population of women were primarily identified from breast cancer records for the

Waikato DHB area included in the New Zealand Cancer Registry (NZCR), for the period from

01/01/1999 to 31/12/2012. Additionally clinic records, operation records, multi-disciplinary

meeting records, oncology, palliative care and other private and public hospital records were

accessed to supplement the NZCR list and also to confirm eligibility of each woman to be

included in this study.

Study eligibility criteria:

1. Newly diagnosed women with primary in-situ or invasive breast cancer between

01/01/1999 and 31/12/2012 (ICD-9 diagnosis codes 174.0 to 174.9, ICD-10 diagnosis

codes 50.0 to 50.9).

2. Morphology consistent with or specific to primary breast cancer originating from the

epithelium of ducts or lobules of the breast. Other varieties of cancer originating from

supporting connecting tissues (i.e., phyllodes tumours and sarcomas) were excluded.

3. Absence of a breast cancer diagnosis prior to 1999 (including the contralateral breast).

4. At the time of the diagnosis, patient residing within the Waikato DHB area or within the

current geographical limits defined by the Waikato DHB

5. Diagnosis of breast cancer made prior to death/ post-mortem

(Note: women who had primary surgery and/or follow up outside Waikato DHB area were

included if all the above criteria were fulfilled)


66

4.2. Data sources:

Data for this study were primarily derived through two sources and were supplemented by

data from several other sources.

For women diagnosed after 01/01/2005, the WBCR was the main data source. For women

diagnosed between 01/01/1999 and 31/12/2004, a retrospective notes review was the main

data source. This review included electronic and hard copy clinical notes from both private

and public hospitals.

4.2.1 The Waikato Breast Cancer Register (WBCR):

The WBCR is a prospectively maintained database that includes all invasive breast cancers in

women who were residents of the Waikato District Health Board area at the time of diagnosis.

The WBCR was established in 2005 and included a process of individual patient consent. The

requirement for patient consent was waived off in 2012. Eligible women with newly

diagnosed breast cancer for the WBCR are identified through clinic records, operation records,

multi-disciplinary meeting records, oncology, palliative care and other private and public

hospital records.

4.2.2 Retrospective data collection:

All women with newly diagnosed cancer between 01/01/1999 and 31/12/2004 were identified

from the NZCR and by accessing clinic records, operation records, multi-disciplinary meeting

records, oncology and palliative care records. All clinical records (i.e., hard copy and

electronic) for these women were accessed and relevant information was extracted into a data

collection form similar to the one used for the WBCR data collection. Subsequently, this

information was transferred into the WBCR database. Further, data for women who were

diagnosed after 2005, but were not included in the WBCR initially, either due to lack of

individual consent (i.e., patient deceased prior to consent) or declined consent were also

collected retrospectively. At the end of this process, a complete dataset of women with newly

diagnosed primary breast cancer between 01/01/1999 to 31/12/2012 was assembled.

Design and Methods

67

4.2.3 Other data sources:

New Zealand Cancer Registry (NZCR):

The NZCR is the national population based cancer registry that records all primary cancers

(excluding non-melanoma skin cancers) in New Zealand. Under the Cancer Registry Act

1993(256), all newly diagnosed cancers are legally required to be reported to the NZCR by the

person in charge of the reporting laboratory. Thus, a copy of the pathology report for each

newly diagnosed cancer is sent electronically to the NZCR which is the major source for new

cancer registrations. Other sources of new cancer registrations include discharge reports from

publicly funded and private hospitals, death certificates and autopsy reports. These non-

histological sources account for less than 10% of all cancer registrations. The Ministry of

Health is responsible for funding and maintaining the NZCR. The NZCR includes a quality

assurance process which is through cross-referencing with data from other population registers

such as the National Minimum Data Set, Mortality Register and other national collections.

Mortality records:

Mortality data were obtained from the National Mortality Database (also known as the

Mortality Collection). The mortality database is maintained by the Ministry of Health and

collects information on all deaths recorded in New Zealand (257). The underlying cause of

death is recorded according to the International Classification of Diseases (ICD) and WHO

Rules and Guidelines for Mortality Coding (109).

Death records with cause specific data were available in the Mortality Collection up to the end

of 2013. For deaths during 1999, cause of death was coded according to the ninth revision of

the International coding of diseases (ICD-9) and for deaths from 2000 onwards, ICD-10

system was used (258).

Pathology Records:

Pathology reports for confirmation of diagnosis and for excised primary tumours was obtained

from patient clinical records, reporting laboratory, or if not available from these two sources,

from the NZCR. For over 98% cancers, a pathology report (i.e., a copy of the authorized

pathology report issued by a pathologist) was available from patient clinical records or from

the reporting laboratory.


68

National Pharmaceutical Database:

Data records for endocrine therapy prescriptions were obtained from the National

Pharmaceutical database for analyses of adherence/compliance with endocrine therapy.

However, these records were found to be adequately complete (i.e., with a completion rate

over 95%) only from 2005 onwards. Therefore these analyses were performed only for women

diagnosed from 2005 onwards, and women who were diagnosed prior to 2005 were excluded.

Although this substantially reduced the sample size and follow up duration, it was deemed to

be the best option to minimize bias, and maximize the quality of the analysis.

National Breast Cancer Screening Database (BreastScreen Aotearoa):

The National Breast Cancer Screening Programme, BreastScreen Aoearoa (BSA) was

established in 1999, and since has collected details of all screen detected and interval cancers

(i.e., cancers diagnosed within 24 months from the last screening mammogram). These data

were used to supplement and to confirm all cases of breast cancer included in the study as

screen detected, interval or non-interval non-screen detected. Further, there were 110 women

who were diagnosed through opportunistic screening mammograms arranged by physicians

outside the BSA programme. These women were also included as screen detected cancers for

analysis.

Design and Methods

69

4.3. Data collection:

4.3.1 Ethics approval:

Ethical approval for this study was obtained from Northern ‘A’ Health and Disability Ethics

Committee (Ethics Ref. 12/NTA/42) to collect breast cancer related information from all

historical cases without individual patient informed consent. Additionally, approval was

obtained from the Kaumatua Kaunihera Research Subcommittee of the Te Puna Oranga

(Māori Health Service) of the Waikato District Health Board.

It was decided not to obtain individual patient consent for this study as we were analysing

historical data from patients diagnosed with breast cancer which will have no influence on

their disease outcomes. Furthermore, this was a large study of which many of the individuals

have already died, but whose information was essential to avoid bias in the study. Practically,

such consent would not be achievable for many and might also have caused significant

difficulties to families of others.

4.3.2 The Waikato Breast Cancer Register (WBCR)

The WBCR has a standardized protocol for identification of eligible women, data collection

and data entry. Eligible women are initially identified at the time of diagnosis from public and

private hospitals in the region. Breast cancer data for these women are collected prospectively

using data collection forms during each step of diagnostic, treatment and follow up pathways.

In addition, all clinical and pathology reports for each eligible woman are accessed and

relevant presenting, diagnostic, treatment (i.e., primary and adjuvant therapy) and follow up

information are extracted into a structured format. Next this information is entered into the

WBCR database which is maintained in a Microsoft Access® database by trained data entry

personnel. Each woman is followed up prospectively through public and private clinic follow

ups and, outcomes including cancer recurrence and death are recorded.

Validity and completeness of the WBCR records are compared annually with breast cancer

records for the Waikato area from the NZCR. Further, data are validated with the Royal

Australian College of Surgeons (RACS) breast section audit data base. Participation in this

audit database and provision of data on breast cancers managed by each surgeon is

compulsory for BSA surgeon accreditation.


70

Quality assurance of the WBCR is maintained through an audit process where at least one in

ten entered records and all records for deceased women are audited by a breast surgeon. At

present, the WBCR is the most complete regional breast cancer register in New Zealand with a

data completion rate of over 98% (259, 260).

4.3.3 Retrospective data collection:

Once eligible women were identified as described above, a process similar to the WBCR was

used to collect breast cancer data. Most of clinical data were extracted through patient medical

notes review, which included pathology reports for all women. Abstracted data from medical

notes review were recorded in a standardized data collection form that was in place for the

WBCR prospective data collection and subsequently transferred into the WBCR database.

Retrospective data collection was performed during 2012 and 2013.

Of the 1131 eligible women identified for this period (1999-2004), breast cancer data for 31

(2.7%) women could not be traced and, hence were excluded from analysis.

4.3.4 Data preparation:

First, data from the retrospective data collection were combined with the WBCR data in its

Microsoft Access database. Follow up information including death and cause of death, local or

metastatic tumour recurrence and whether free of disease at last known follow up was also

recorded. Next, all relevant data from this database were exported into a Microsoft Excel

datasheet where data cleaning was undertaken. This procedure created a dataset with a single

observation for each patient or for each cancer episode for women with metachronous cancer

(i.e., second new breast cancer) during the study period. Additional linking of this completed

dataset with data from the NZCR (for stage comparison and validation), National

Pharmaceutical database (for adherence to endocrine therapy), BreastScreen Aotearoa (for

comparison of differences in screen detected versus non-screen detected cancer) and National

Minimum Dataset (for comorbidities) were done in Microsoft Excel datasheet using patient

NHI and date of diagnosis.

Once cleaning, integration and data linkage were completed, the dataset was imported to SPSS

Version 22 where all statistical analyses were performed (261).

Design and Methods

71

4.4. Variables:

4.4.1 Exposure variable - Ethnicity:

Ethnicities (or Race) are constructed categories that may reflect history, geographic origin,

cultural identity, socioeconomic status and sometimes genetics and biology, all in varying

degrees. Human geneticists and anthropologists agree that pure human races do not exist

(262). However, internal and external racial/ethnic identifications allow races/ethnicities to

exist and sustain as social constructs (263). Irrespective of their origins or basis of formations,

different racial/ethnic groups have substantially different rates of diagnosis, treatment, and

outcome in a variety of diseases, including cancer in many different countries (4, 10, 11).

According to the definition used by New Zealand Statistics, ethnicity is the ethnic group or

groups that people identify with or feel they belong to, which is a measure of cultural

affiliation, as opposed to race, ancestry, nationality or citizenship. Using this definition,

ethnicity is seen as self-perceived and people can belong to more than one ethnic group (264).

Although the terms race and ethnicity are sometimes used interchangeably, in many instances

they have different meanings. In general, ethnicity refers to shared cultural practices,

perspectives, and distinctions that set apart one group of people from another. However, the

definition of ethnicity differs by country. For example, in neighbouring Australia ethnicity is

defined as relating to or peculiar to a human population or group, especially one with a

common ancestry, language, etc., or relating to the origin, classification, characteristics, etc.,

of such groups.

In New Zealand context, Māori and NZ European are recognized as two ethnic, but not racial

groups. In fact many who identify themselves as Māori have a significant European

inheritance. Therefore this thesis refers to Māori, European and other groups included in this

study as ethnic and not as racial groups.

Up to the population census in 1981 Māori were defined as those with half or more Māori

ancestry. But from the population census in 1986, the question relating to the ethnic origin

allowed people to identify themselves using one or more of the following options: European,

Māori, Pacific, Chinese, Indian or other ethnic origin irrespective of their ancestry.

Three approaches for assigning Māori ethnicity in health research are documented. First

approach is to assign Māori ethnicity based on documented or self-identified ethnicity at a

given point of time. This method has been shown to be associated with high risk of


72

misclassification leading to a significant undercounting for Māori. Second method, the ‘ever-

Māori’ approach has the least likelihood of undercounting, but carries the risk of over-

counting Māori especially for women who had multiple episodes of contact with health care

system. This method has been shown to provide ethnicity distributions close to New Zealand

Census Mortality Study (NZCMS) adjusted estimates for epidemiological studies, which is

considered to be the gold standard (133). A third approach, that is ‘ever-Māori’ up to a given

point in time balances biases of those two approaches as it minimizes risks for both over and

under-counting.

Ethnicity was self-assigned by the patients for prospective data included in the WBCR; each

woman was requested to fill the ethnicity field in the data collection form during consent

process. For retrospective data collection (pre 2004) ethnicity data were obtained from patient

clinical records maintained in electronic and hard copy formats, as per the Ministry of Health

ethnicity data protocols (265).

For patients included in the prospective database, self-identified ethnicity was considered as

the final ethnicity, while for retrospectively data collected women, Ever-Māori at a given point

in time approach were used. With this approach, if any health service document has identified

the woman of interest as Māori up to the point of diagnosis, she was assigned Māori ethnicity.

Ethnicity was recorded in the database using standard ethnicity codes from New Zealand

Statistics. For patients with more than one recorded ethnicity, a prioritization system was used

to assign ethnicity, giving highest priority to Māori followed by Pacific, Asian, Other (except

NZ European) and NZ European, based on New Zealand Statistics ethnicity classification

system. Subsequently, ethnicity was grouped by four categories for analysis; Māori, Pacific

(including Samoan, Cook Island Māori, Tongan, Niuean, Tokelauan, Fijian and other pacific

island), NZ European (European not further defined, NZ European, other European) and Other

(Asian not further defined, South East Asian, Indian, Chinese, other Asian, Middle Eastern,

Latin American / Hispanic, African and other).

Design and Methods

73

Table 3: Definitions of socio-demographic, tumour and treatment characteristics

Characteristic Variable Values Comments / Definitions

Ethnicity Ethnicity Māori

NZ European

Pacific

Other

For women diagnosed after 2005 - Self-identified ethnicity

from the WBCR

For women diagnosed during 1999-2004 - Self-identified

ethnicity recorded in health service records up to the time of

diagnosis using “Ever-Māori” approach

Time origin Date of diagnosis From 01/01/1999 to 31/12/2012 Date of obtaining tissue (e.g. core biopsy)/cells (e.g. FNAC)

from the primary tumour or secondary deposits that

confirmed the diagnosis of breast cancer

Demographics Age at diagnosis 20-99 years Based on date of birth and date of diagnosis

Year of diagnosis 1999-2012 Based on date of diagnosis

Tumour

characteristics

Stage at diagnosis TNM system – stages 0 and I to IV

SEER– localized, regional & distant

Tumour size 0-210 mm Maximum tumour diameter based on histopathology report

Number of positive nodes 0-48 Total number of regional lymph nodes found to be invaded

by tumour (macro or micro metastases, but excluding lymph

nodes with only isolated tumour cells) based on

histopathology report

Tumour grade Grade I-III Based on the Elston and Ellis modified Scarff-Bloom-

Richardson breast cancer grading system as reported in the

histopathology report

Histopathology type Ductal, Lobular, Mixed, Other Derived from histopathology report in accordance with

World Health Organization Classification of tumours of the

breast


74


Hormone receptor status -

Oestrogen (ER) &

Progesterone (PR)

Positive / Negative Based on the results of immunohistochemistry tests as

reported in histopathology report

HER-2 status Immunohistochemistry – Positive,

Equivocal or negative

Fluorescent in-situ hybridization –

Positive or Negative

Based on Fluorescent In-Situ Hybridization (FISH) test or

when this was not available, on immunohistochemistry

Lympho-vascular invasion

(LVI)

Positive or negative Based on histopathology report

Patient

characteristics

Comorbidity score 0-8 According to the Charlson Comorbidity Index based on

documented comorbidities at the time of diagnosis

Body mass index (BMI) 1.8-59.9 Calculated using body weight and height where available

Smoking status Non, ex-smoker or current smoker As documented at the time of diagnosis

Diagnosis Mode of detection Screen detected

Non-screen detected

Interval cancer

Screen - detected through a screening mammogram

performed on an asymptomatic woman

Non-screen - cancer diagnosis process initiated following

symptoms directly or indirectly related to the breast cancer

Interval - if diagnosis within 24 months from last screening

mammogram

Health care access Small area deprivation 1-10 Based on NZ Deprivation Index 2006. Deprivation deciles

were aggregated to form quintiles (1 to 5) for analysis

Residence Urban, semi-urban, rural Based on NZ Statistics Urban/Rural classification

Distance from hospital 0-25, 25-50, 50-100, >100km Distance from residence to tertiary hospital in Hamilton

Facility type Public or private Facility type where primary cancer surgery was performed

Design and Methods

75


Surgical treatment Type of primary operation to

the breast

Mastectomy, breast conserving

surgery (BCS) or no primary surgery

Mastectomy - complete surgical removal of the ipsilateral

breast, BCS - any operation for the breast cancer which was

less than a mastectomy, No primary surgery - When no

primary surgical excision done due to patient not fit for an

operation, advanced nature of cancer or due to patient

declining

If primary mastectomy,

decision taken by

patient / surgeon Based on clinical records on decision making process for

mastectomy. This information was clearly available for only

79.9% of women who underwent primary mastectomy

Type of operation to the axilla Sentinel node biopsy (SNB) based

management, primary axillary node

dissection (ALND), axillary

sampling or no axillary surgery

SNB – if a radio isotope or blue dye or both have been used

to ascertain nodes draining the breast/tumour. ALND – If

standard level II or III dissection been done, Node sampling –

where only enlarged discrete nodes were removed without a

formal ALND

Re-excision Yes / No When a secondary wider excisional surgery is performed on

the ipsilateral breast following a BCS to clear any residual

tumour and to achieve a cancer free margin

Breast reconstruction Yes / No Major breast reconstruction (immediate or delayed)

performed following a total mastectomy

Timeliness of surgery 0-364 Time gap from the date patient informed of the diagnosis of

breast cancer to the date of primary surgical operation

Adjuvant

Chemotherapy

Received chemotherapy Yes / No If the patient received a single or more doses of

chemotherapy prior to or after surgery

Timeliness of chemotherapy 10-232 days Time gap in days from most invasive surgery for cancer to

initiation of chemotherapy


76


Adjuvant Radiation

therapy

Received radiotherapy Yes / No If the patient received a single or more fractions of

radiotherapy to the breast or chest wall prior to or after

surgery

Timeliness of radiotherapy 17-364 days Time gap from most invasive surgery for women not

receiving chemotherapy and for women receiving

chemotherapy time gap from chemotherapy completion to

radiotherapy

Adjuvant Endocrine

therapy

Received endocrine therapy Yes / No If the patient has obtained endocrine therapy from a

pharmacy after the date of diagnosis based on Pharmaceutical

Database records

Adherence to endocrine

therapy

Good / Sub-optimal Good – if medication possession ratio (adherence index) was

≥80% over the follow up period up to 5-years or up to death,

whichever was shorter.

Sub-optimal – if medication possession ratio was <80%

Biological therapy Received biological therapy Yes / No If a woman has received adjuvant biological therapy

(Trastuzumab) for a HER-2 amplified cancer

Design and Methods

77

4.4.2 Tumour characteristics

Stage at diagnosis

Many different staging systems exist for staging of breast cancer. Tumour, Node and

Metastasis (TNM) system and the Surveillance Epidemiology and End Results (SEER)

program cancer staging definitions are the commonly used of these systems (266, 267). This

study primarily used the TNM staging system which is used by a majority of clinicians and

was also used in many similar studies, which provided an opportunity for comparison. The

SEER system was used in situations where comparisons were performed with the NZCR

which primarily uses the SEER system.

Cancer stage was ascertained based on all available clinical, imaging (within four months

before or after date of diagnosis) and pathology information from tumour excision.

TNM staging:

TNM system published by the International Union Against Cancer (UICC) and the American

Joint Committee on Cancer (AJCC) provides a universal approach to staging of cancer. The

TNM system uses status of primary tumour (T), lymph node involvement (N) and metastatic

deposits (M) to ascertain cancer stage. TNM summary stages (stages I to IV) are formed based

on different TNM combinations (Table 4).

Table 4: TNM staging system for staging of breast cancer

TNM stage Description

Tumour (T)

Tis Carcinoma in-situ

T1 Invasive tumour 20mm or less in greatest dimension

T2 Invasive tumour more than 20mm but less than 50mm in greatest dimension

T3 Invasive tumour more than 50mm in greatest dimension

T4 Invasive tumour of any size with direct extension to the chest wall and/or to the

skin (ulceration or skin nodules).


78

Nodes (N)

N0 Regional nodes not involved

N1

Micro-metastases or metastases in 1–3 axillary lymph nodes; and/or in internal

mammary nodes with metastases detected by sentinel lymph node biopsy but not

clinically detected

N2 Metastases in 4–9 axillary lymph nodes; or in clinically detected internal

mammary lymph nodes in the absence of axillary lymph node metastases

N3

Metastases in ten or more axillary lymph nodes; or in infra-clavicular (axillary

level III) lymph nodes; or in clinically detected ipsilateral internal mammary

lymph nodes in the presence of one or more positive level I, II axillary lymph

nodes; or in >3 axillary lymph nodes and in internal mammary lymph nodes with

micro-metastases or macro-metastases detected by sentinel lymph node biopsy but

not clinically detected; or in ipsilateral supraclavicular lymph nodes

Metastasis (M)

M0 No clinical or radiographic evidence of distant metastases

M1 Distant detectable metastases as determined by classic clinical and radiographic

means and/or histologically proven larger than 0.2 mm

Stage grouping

Stage I T1N0M0, T1N1micM0

Stage II T1N1M0, T2N0M0, T2N1M0, T3N0M0

Stage III T1-2N2-3M0, T3N1-3M0, T4N0-3M0

Stage IV Any T, Any N, M1

The TNM staging is generally performed at two points for a majority of breast cancers. First is

done (clinical TNM / cTNM) once the diagnosis is confirmed prior to treatment. Second or

pathological TNM staging (pTNM) is done based on histopathology report from the primary

tumour excision. Additionally in situations where a patient undergoes neo-adjuvant therapy

(up-front chemotherapy or endocrine therapy prior to surgical excision) clinical staging is

performed for a second time prior to surgery and is denoted as yTNM.

Design and Methods

79

For this study, where ever it was available, pTNM stage was considered as the final tumour

stage. In situations where pTNM was unavailable; for example in situations where primary

surgery was not performed, or if a patient has received neo-adjuvant therapy, cTNM was

considered as the final cancer stage. For prospective data collection AJCC/TNM 7th version

was used and for retrospective data (pre 2004) 6th version was used.

SEER cancer stage (Extent of disease):

The SEER program cancer staging definitions (Table 5) are published by the United States

National Cancer Institute and are preferred by many cancer registries for cancer stage

recording including the NZCR, due to its simplicity (266). This system summarized cancers

into localized, regional or distant.

The TNM and the SEER system are neither comparable nor interchangeable. For example,

invasion into pectoralis major muscle is considered as regional disease in the SEER system

while the TNM system does not acknowledge this fact in its classification.

The NZCR primarily uses the SEER program cancer staging definitions published by the

National Cancer Institute of the USA (266). For each reported case of cancer to the NZCR,

stage is manually determined by experienced cancer coders, primarily using pathology report

from the primary tumour excision together with additional information from hospitalization

records, death certificates and autopsy reports. Stage is assigned for each cancer based on

staging data available at the end of the first course of therapy, or within four months of the

date of diagnosis, whichever is earlier (268).

Table 5: SEER summary staging system for staging of invasive breast cancer

SEER summary stage Description Equivalent TNM stages

Localized Invasive tumour confined to the breast T1-4 N0 M0

Regional Invasive tumour invading to adjacent tissue

(i.e. skin or chest wall including pectoral

muscle a) or regional lymph nodes

T4 N0 M0, Any T N1-3

M0

Distant Tumour with distant metastasis Any T, Any N, M1

a Pectoral muscle invasion not recognized as local invasion in the TNM system


80

Cancer grade:

Invasive tumour grade was defined according to the Elston and Ellis modified Scarff-Bloom-

Richardson breast cancer grading system (269). Tumour grades reported as; well, moderately

or poorly differentiated were considered as corresponding to Scarff-Bloom-Richardson grades

1, 2, and 3 respectively. As tumour grading requires a histological specimen, women for

whom the diagnosis was confirmed only on cytological features from a FNAC and has not had

further biopsy, grading was unavailable. There were 201 (7.1%) women for whom the cancer

grade was not available due to this reason. Significant proportions of these women had either

advanced cancers or were deemed to be too ill for further investigations or treatment.

Histopathology:

Invasive or in-situ cancers arising from either ducts or lobules of the breast only were included

and histologic types of the tumours were recorded in accordance with World Health

Organization Classification of tumours of the breast. Histopathology was classified into ductal,

lobular, mixed or other for analysis. Histopathology type was unavailable for women on whom

the diagnosis was made based only on cytology (FNAC).

Hormone receptor status:

Oestrogen (ER) and progesterone (PR) receptor status was determined based on the results of

immunohistochemistry tests and classified as positive or negative. ER status was missing from

60 (2.1%) and PR status was missing from 113 (4%) women with invasive cancer.

Human Epidermal Growth Factor Receptor type-2 (HER-2) status:

HER-2 status was based on Fluorescent In-Situ Hybridization (FISH) test or when this was not

available, on immunohistochemistry (270). Immunohistochemistry results were available as

positive, equivocal or negative while FISH reports which are more specific report as HER-2

positive or negative.

Trastuzumab (Herceptin) for adjuvant treatment of HER-2 amplified early breast cancer

became fully funded through the public health system in New Zealand in 2007 (61), and

assessment of HER-2 status has only been routine in New Zealand since then. This has

resulted in a relatively high rate of missing HER-2 data in our study, especially among women

Design and Methods

81

diagnosed with breast cancer prior to 2007. Although the overall missing HER-2 rate was

26.2%, missing rate was only 4.3% among women diagnosed after 2006.

Lympho-vascular invasion:

Lympho-vascular invasion (LVI) refers to invasion of lymphatic spaces, blood vessels, or both

in the peri-tumoural area by tumour emboli which are critical steps in metastasis. LVI is

routinely reported for all invasive cancers where tissue was available for histopathological

analysis. Hence, similar to grade and histology type, this was missing for women for whom

the diagnosis was made only on cytology from FNAC.

4.4.3 Patient characteristics:

Comorbidities:

Any significant coexisting medical conditions present or detected at the time of breast cancer

diagnosis were considered as comorbid conditions. Comorbid conditions were identified from

review of medical and anaesthetic records, multi-disciplinary team meeting records and from

oncology records. This was further supplemented with comorbidity data from the the National

Minimum Dataset (NMDS) which include records of comorbid conditions documented during

all public and private hospital inpatient episodes. Past conditions that have completely

resolved and have no known significant long term consequences (e.g. appendicectomy,

cholecystectomy) were excluded.

The CCI uses 19 medical conditions, each allocated a weight of 1 to 6 depending on the

adjusted relative risk of 1-year mortality, and added together to give an overall score. The CCI

has been validated in a cohort of breast cancer patients with the 10-year mortality rate as an

endpoint (152). The CCI score was categorized into 0, 1-2 and 3+ for analysis.

(A group of researchers led by Prof. Diana Sarfati at the University of Otago is conducting a

nationwide study (C3 study) which is aimed at identifying the impact of comorbidity on ethnic

disparities in cancer outcomes in New Zealand. The WBCR has also contributed its data to this study

which will publish its final findings in the near future. In light of this large ongoing study, impact of

comorbidities is not studied in great detail in this thesis, although comorbidity is considered as an

important covariate and adjustments done accordingly.)


82

Smoking status:

Smoking status was denoted as non-smoker, ex-smoker or current smoker based on records

from the WBCR (patient declared during consent process) or from medical notes review for

retrospectively data collected women. Smoking data were missing for 217 (7.6%) women with

invasive cancer.

For analysis including survival modelling, missing smoking status was considered as a

separate categorical variable. A separate sensitivity analysis was also performed using imputed

smoking data for these missing cases. Imputation was done using variables including age,

ethnicity, residence status (urban/rural), deprivation and comorbidity. This analysis yielded

results much similar to the first analysis where a separate missing category was used.

Survival analysis was repeated using only cases with complete data for all variables and the

results of this analysis was also found to be almost similar to the full dataset.

Body mass index (BMI):

Weight and height data for each woman was documented as measured at or closest to the time

of diagnosis. BMI was calculated by dividing weight in kilograms by square of height in

meters. Both height and weight data were available only for 72.2% women with invasive

breast cancer. Missing BMI was disproportionately higher for women not undergoing primary

surgery (40.5%). For analysis, missing BMI status was considered as a separate categorical

variable. However a selection bias is likely and hence caution is needed in interpreting these

results.

Design and Methods

83

4.4.4 Health care access

Health care access parameters were determined based on individual patient’s address at the

time of diagnosis. This was used to assign area level socioeconomic deprivation, urban/rural

residential status and to determine distance from the health care facility.

Area level deprivation

Socioeconomic deprivation status of each woman was determined using the New Zealand

Deprivation Index 2006 (NZDep06) (271). The NZDep06 measures deprivation level based on

place of residence at the time of cancer diagnosis. This index uses nine variables (benefit

income, employment, household income, communication, transport, support, qualifications,

living space, and home ownership) as measured during the 2006 national census. The

NZDep2006 has created small areas (mesh-blocks covering a population of approximately

100) in a geographical map, on a deprivation scale from 1 to 10; 10 represents the most

deprived 10% of New Zealand areas, while 1 represents the least deprived 10% of areas.

Each patient’s physical address at the time of diagnosis was used to assign each patient into an

area of deprivation. For comparisons each two consecutive deciles were combined together to

form five categories i.e. 1 and 2 – least deprived, 9 and 10 – most deprived.

A small number of area units are not included in the New Zealand Statistics concordance file.

For women from these areas a deprivation decile was derived using the same method as used

by the New Zealand Statistics. That is, for each census area in question, the population-

weighted average NZDep score was calculated from all the mesh blocks that made up that

area, and this NZDep score was then used to allocate the appropriate NZDep2006 decile.

Residential status:

Urban / rural residential status was determined based on individual residential address

according to the Statistics New Zealand’s urban rural classification system (Figure 13). This is

a seven-level classification which allocates each census area based on both population size and

mobility or access to urban amenities and services.

Main urban areas represent the most urbanised areas in New Zealand. Main urban areas are

very large and centred on a city or main urban centre. They have a minimum population of


84

30,000. Population size is also used to define secondary and minor urban areas in the standard

urban area classification. But population size alone cannot adequately describe the

characteristics of different urban areas. Satellite urban category identifies towns and

settlements with strong links to main urban centres and independent urban category identifies

towns and settlements without significant dependence on main urban centres.

Rural areas are the other remaining areas of the country where the population does not exceed

999. Rural areas with a high urban influence category identify rural areas that form a transition

between the main urban areas and rural areas, although mesh blocks are not necessarily

contiguous with main urban centres. Rural areas with moderate urban influence category

identify rural areas with a significant, but not exclusively, main urban area influence while

areas with a strong rural focus are categorized as rural areas with low urban influence with

majority of the population working within the same rural area. Highly rural or remote are rural

areas where there is minimal dependence on urban areas in terms of employment, or where

there is a very small employed population.

These seven levels were categorized into a three levels; urban category included main and

satellite urban communities. Semi-urban/semi-rural category included independent urban

communities and rural areas with high or moderate urban influence. Rural category included

rural areas with low urban influence and highly rural areas.

Figure 13: New Zealand Statistics – Urban/Rural Classification system (Source NZ Statistics)

Urban Semi-urban/semi-rural Rural

Design and Methods

85

Distance from treatment centre:

Distance from treatment centre was based upon individual patient address. As almost all

women included in this study have received surgery, and adjuvant radiation and chemotherapy

from the tertiary public hospital and/or private treatment facilities in Hamilton, distance was

calculated from patient address to Hamilton city. Distances were categorized into <10km, 10-

50km, 50-100km and >100km (129).

Treatment facility type:

Treatment facility type was categorized into public or private based on facility where patient

underwent primary surgical treatment. As radiation therapy for all women and chemotherapy

for almost all women (except for a few receiving chemotherapy in private in Auckland) were

provided from the public sector, facility type was not included as a variable for analysis.

However, we recognized the differences in socioeconomic, educational and behavioural

characteristics of women receiving surgery in private sector and hence this was included as a

variable in assessment of time delay and coverage of adjuvant radiation and chemotherapy.

Treatment of breast cancer:

Details of most of treatment variables include in analyses are shown in Table 3. For primary

surgical therapy type and timeliness of therapy for breast and axilla were documented. A

diagnosis to treatment gap of 31 days was considered as threshold delay to assess timeliness of

surgery as described in Faster Cancer Treatment Indicators (Figure 14) published by the

Ministry of Health (58). For adjuvant radiation and chemotherapy, coverage and timeliness

were documented. Threshold limits of 60 days for chemotherapy (from date of definitive

surgery) and 90 days for radiotherapy (from date of definitive surgery for women not

undergoing chemotherapy) were used based on published literature.


86

Figure 14: Faster Cancer Treatment Indicators 2012-2013 (Source - Ministry of Health)

For adjuvant hormone therapy coverage and adherence with endocrine therapy was calculated

based on data from the Pharmaceutical database. Coverage of adjuvant therapy was analysed

based on accepted treatment guidelines over the period of the study to identify women who

should have received respective types of adjuvant therapies (44, 272).

Urgent referral

with high-

suspicion of

cancer

First cancer

treatment

First specialist

assessmentDecision-to-treat

Indicator one (best practise – 62 days)

Indicator two (best practise – 14 days) Indicator three (best practise – 31 days)

Design and Methods

87

4.5. Outcome data

Primary outcome of interest in survival analysis was death (due to breast cancer or other

causes) or survival. Details and dates of tumour recurrences were included for disease free

survival analysis. Cause and date of death was identified from patient clinical records and

from the Mortality Collection while tumour recurrence data were based on patient clinical

records. End point for patient follow up was death or last follow up recorded up to 31/12/2013.

Table 6: ICD-9 and ICD-10 codes corresponding to breast cancer as underlying cause of death

Year of

death

Coding

system

Codes representing death due to breast cancer

1999 ICD-9-CM 1740, 1741,1742, 1743, 1744, 1745, 1746, 1747, 1748, 1749

1750, 1751, 1752, 1753, 1754, 1755, 1756, 1757, 1758, 1759

2000-2013 ICD-10-CM C500, C501, C502, C503, C504, C505, C506, C507, C508, C509

The Mortality collection assigns cause of death codes according to the World Health

Organization rules and guidelines for mortality coding (Table 6). This is done based on

information received from death certificates, post-mortem records, hospital discharge records

and Coroners reports.

In general there was a good concordance (>90%) between cause of death derived from clinical

notes review (where details of death or events preceding death were documented) and cause of

death from the Mortality collection.


88

4.6. Sample size estimation

This study was estimated to include approximately 3,000 women (450 Māori and 2,550 NZ

European) with breast cancer. It was estimated that 5-year breast cancer specific survival

during the study period to be approximately 78% for Māori and Pacific women and 84% for

NZ European women (9, 273). To show that a variable of interest results in a 5% worse 5-year

survival amongst Māori (2-tailed p value, 95% confidence level and 80% statistical power) it

was required to include 430 Māori and 2,170 NZ European women.

4.7. Data analyses

Different methods of analyses were applied for analyses given under different sub-chapters in

the results section. Basic principles of analyses are described here and details of analyses

including specific statistical tests applied for each study of interest are described in each

results section.

Analyses under each section are broadly categorized into three areas. First, Māori and NZ

European cohorts were compared in terms of demographics, disease and patient characteristics

and healthcare access factors. Second, treatment characteristics were compared between Māori

and NZ European women adjusting for age, tumour characteristics and health care access

characteristics. Third, the impact of each characteristic/factor of interest on breast cancer

mortality/survival was compared adjusting for covariates including ethnicity (Māori versus NZ

European). In the final survival analysis breast cancer mortality hazard ratios were

sequentially adjusted for demographics, health care access, tumour biological characteristics,

comorbidity and treatment factors, to identify the overall quantitative impact of each of these

factors on observed mortality inequity.

All analyses were performed in SPSS version 22 (261).

Results

89

Chapter 5. Results

This chapter presents results from data analysis. It is set out in different sections. First section

analyses the validity of data followed by eight sections that examine different areas and causes

of breast cancer inequities between Māori and NZ European women. The final section is a

combined analysis of results from previous eight sections which aims to create a model to

describe quantitative impacts of patient, cancer and healthcare service factors on the survival

disparity between Māori and NZ European women.

Each of the ten sections included are either published papers or papers that have been

submitted for publication. These papers have been modified and abbreviated from their

original versions to ensure segue between sections and to minimize repetitions. General

changes to each section include shortening of introduction limiting them to study question/s,

methods section limiting only to methods specific to each study and modifications to the

discussion sections to make them more succinct and to minimize repetition.

Some of the studies included contain different sample sizes. This is either due to respective

study been done prior to completion of full data collection or due to limitations of data

availability from other sources which were used for data linkage with the WBCR (e.g.

National Pharmaceutical Database).

Each section starts with an abstract and a short introduction to the topic with the study

question followed by methods specific to the study and results. Discussion in each section is

limited to discussing specific issues raised during each study. Next chapter (Chapter 6)

provides a summary discussion which is aimed at summarizing all study findings and to bring

together these findings for final interpretations.


90

Results

91

5.1. How valid are the data used in this study?

Preface:

This chapter contains an abbreviated version of a manuscript published in Cancer

Epidemiology.

Authors: Seneviratne S, Campbell I, Scott N, Shirley R, Peni T, Lawrenson R.

Title: Accuracy and completeness of the New Zealand Cancer Registry for staging of

invasive breast cancer

Journal: Cancer Epidemiology

Year of publication: 2014

DOI: 10.1016/j.canep.2014.06.008

Impact factor: 2.56

Journal’s aims and scope: This journal is dedicated to increasing understanding about

cancer causes, prevention and control. The scope of the journal embraces all aspects of

cancer epidemiology including: descriptive epidemiology and statistics, studies of risk

factors for disease initiation, development and prognosis, screening, early detection

and accurate diagnosis, prevention and evaluation of interventions and methodological

issues and theory.


92

Abstract:

Background:

Population based cancer registries are an invaluable resource for monitoring incidence and

mortality for many types of cancer. Research and healthcare decisions based on cancer registry

data rely on the case completeness and accuracy of recorded data. This study was aimed at

assessing completeness and accuracy of breast cancer staging data in the New Zealand Cancer

Registry (NZCR) against the Waikato Breast Cancer Register (WBCR).

Methods:

Data from 2562 women diagnosed with invasive primary breast cancer between 1999 and

2011 included in the WBCR were used to audit data held on the same individuals by the

NZCR. WBCR data were treated as the benchmark.

Results:

Of 2562 cancers, 315 (12.3%) were unstaged in the NZCR. For cancers with a known stage in

the NZCR, staging accuracy was 94.4%. Lower staging accuracies of 74% and 84% were

noted for metastatic and locally invasive (involving skin or chest wall) cancers respectively,

compared with localized (97%) and lymph node positive (94%) cancers. Older age (>80

years), not undergoing therapeutic surgery and higher comorbidity score were significantly

(p<0.01) associated with unstaged cancer. The high proportion of unstaged cancer in the

NZCR was noted to have led to an underestimation of the true incidence of metastatic breast

cancer by 21%. Underestimation of metastatic cancer was greater for Māori (29.5%) than for

NZ European (20.6%) women. Overall 5-year survival rate for unstaged cancer (NZCR) was

56.3% which was worse than the 5-year survival rate for regional (78.8%), but better than

metastatic (18.2%) disease.

Conclusions:

Unstaged cancer and accuracy of cancer staging in the NZCR are major sources of bias for the

NZCR based research. Improving completeness and accuracy of staging data and increasing

the rate of TNM cancer stage recording are identified as priorities for strengthening the

usefulness of the NZCR.

Results

93

Background:

Population based cancer registries are a valuable resource for monitoring incidence and

mortality from cancer and play a vital role in cancer control programmes (274). Many national

cancer control strategies including the New Zealand Cancer Control Strategy have recognized

the importance of a high quality national cancer registry as a core component of cancer control

(275).

According to the World Health Organization, a modern cancer registry is expected to provide

data on a number of key areas (274). These include enabling the assessment of the current

magnitude of the cancer burden and future projections, providing a basis for research on

cancer causes and prevention, providing information on prevalence of risk factors, and

monitoring the effects of prevention, screening, treatment and palliative care. Quality of a

cancer registry forms a cornerstone from which to achieve these tasks. The International

Agency for Research on Cancer describes five main components of quality for cancer

registries (276). These include completeness in cover, completeness in detail, accuracy in

detail, accuracy of reporting and accuracy of interpretation.

Several studies have raised the issue that substantial proportions of cancers are unstaged or

staged inaccurately in the NZCR (16, 94, 277). For example an audit on colon cancer by

Cunningham and colleagues reported a staging accuracy of 80% in the NZCR compared with

stage determined from a clinical notes review (16). Another audit comparing lung cancer

staging in the NZCR against a regional database reported a staging accuracy of only 43.8%

(94). The same audit reported that 12% of cases out of 565 included were not known to the

NZCR. Missing or inaccurate cancer stage data may lead to biased research results. A good

understanding of completeness, accuracy and characteristics associated with unstaged cancer

in the NZCR is required to understand the magnitude of bias and will enable rational

conclusions to be drawn from cancer research.

The Surveillance Epidemiology and End Results (SEER) program cancer staging definitions

are preferred by many cancer registries for cancer stage recording (266), including the NZCR

due to its simplicity. However, clinicians and pathologists widely use the Tumour Node

Metastases (TNM) staging system which is more detailed and more relevant for clinical

decision making (267). Since its introduction to the NZCR in 2001, TNM stage recording has

slowly been increasing and was approximately 50% complete for breast cancer in 2010

(personal communication with the NZCR) which is well below the SEER stage completion


94

rate in the NZCR over this period (278). Comparatively, cancer registries from countries such

as Denmark and the Netherlands have achieved TNM completion rates of more than 90% for

many cancers including breast cancer (279, 280).

This study was conducted to evaluate the completeness and accuracy of breast cancer data

from the NZCR against the WBCR. Further analyses were done to identify patient

characteristics associated with unstaged cancer and to compare outcome for unstaged against

staged cancers. Details of cancers from 2012 were not available at the time of this study, and

hence only women diagnosed between 1999 and 2011 were included in this study.

Methods:

Data:

All newly diagnosed primary invasive breast cancer records over a 13-year period from

01/01/1999 to 31/12/2011 were identified from the WBCR and compared with the same

records for the Waikato District Health Board (DHB) area from the NZCR. Each record was

matched by date of diagnosis and National Health Index (NHI) number. From a total of 2623

invasive breast cancers identified for the period under review from the WBCR and the NZCR,

women with a post mortem diagnosis of breast cancer (n=4) and women recorded under a

different area (n=9) were excluded. Four cases from the WBCR not known to the NZCR and

43 cases from the NZCR not included in the WBCR (ineligible due to residence outside

Waikato DHB area or due to records not available to the WBCR) were excluded from

comparisons.

Variables:

Extent of disease (i.e. degree of spread of the tumour within the body / tumour stage) from the

NZCR for selected breast cancers were compared with same data from the WBCR. WBCR

data were treated as the benchmark and completeness and accuracy of the NZCR staging

records were analysed against the WBCR.

All stage comparisons were performed according to SEER extent of disease classification

which classifies tumours into 4 categories; localized, locally invasive (into skin or chest wall),

involving regional lymph nodes (LN) and metastatic (i.e. tumour spread beyond breast and

regional lymph node) disease (266).

Results

95

As a pathology report from primary tumour excision would only be available for women

undergoing primary surgical interventions, primary therapeutic surgical intervention was

included as a predictor for staged cancer in the NZCR. Women undergoing only diagnostic or

palliative surgical interventions or undergoing no surgical treatment were classified as no

therapeutic surgical intervention.

Statistical analysis:

WBCR data for all women with an unknown stage in the NZCR were explored in a univariate

analysis using Chi squared (χ²) tests for trend. For staged cancers in the NZCR, sensitivity and

specificity for each stage were calculated against the WBCR stage. Factors associated with

unstaged cancer were explored in a multivariable logistic regression model. Survival for each

cancer stage in the NZCR and the WBCR were compared using Kaplan-Meier survival curves.

A Cox proportional hazard model was used to estimate the risk of mortality for unstaged

compared with staged cancers adjusting for age, comorbidity, ethnicity and cancer stage.

Results:

From a total of 2623 newly diagnosed primary invasive breast cancers identified from the

WBCR and NZCR, 2563 cancers were found to be eligible for this study. Of these, 1 cancer

for which the stage was not recorded in the WBCR was excluded leaving 2562 cancers for

stage comparison.

Table 7 shows the distribution of discrepancies between the WBCR and the NZCR in relation

to extent of disease. Overall, 315 (12.3%) of cancers in the NZCR were recorded as unknown

stage. Of the cancers with a known stage in the NZCR, 2121 (94.4%) were found to be

accurately staged compared with the WBCR. Higher proportions of unstaged and inaccurately

staged cancers were seen for metastatic and locally invasive cancers compared to localized and

lymph node positive cancers. Sensitivity of each extent of disease category in the NZCR was

97.7%, 84%, 93.7% and 73.9% for localized, locally invasive, LN involved and metastatic

cancer respectively. Specificities for respective extents of disease were 95.3%, 99.5%, 96.5

and 99.4% (Figure 15).


96

Table 7: Extent of cancer (stage) at diagnosis in the New Zealand Cancer Registry compared with

extent of cancer at diagnosis in the Waikato Breast Cancer Register 1999-2011.

WBCR extent of cancer

NZCR extent of

cancer

Localized Locally invasive LN involved Metastatic Total

n % n % n % n % n %

Localized 1186 84.2 4 10.3 44 4.6 1 0.7 1235 48.2

Locally invasive 4 0.3 21 53.8 4 0.4 2 1.3 31 1.2

LN involved 20 1.4 0 0.0 837 86.8 27 17.9 884 34.5

Metastatic 4 0.3 0 0.0 8 0.8 85 56.3 97 3.8

Unknown 194 13.8 14 35.9 71 7.4 36 23.8 315 12.3

Total 1408 100 39 100 964 100 151 100 2562 100

Stage distribution for staged cancers in the NZCR was highly and significantly (p<0.001)

correlated with the overall stage distribution in the WBCR over the study period (Figure 16).

The highest correlation was observed for locally invasive cancers (correlation

coefficient=0.81), while the correlations of 0.75 and 0.69 were observed for regional and

metastatic cancer, respectively.

Figure 15: Distribution of accurately staged, inaccurately staged and unstaged breast cancer in the

New Zealand Cancer Registry compared with the Waikato Breast Cancer Register 1999-2011.

83.7%

53.8%

86.8%

56.3%

2.6%

10.3%5.8%

19.9%

13.8%

35.9%

7.4%

23.8%

0%

20%

40%

60%

80%

100%

Localized(n=1408)

Locally invasive(n=39)

LN involved(n=964)

Metastatic(n=151)

Accurately staged

Inaccurately staged

Unstaged

Results

97

Figure 16: Trends in proportional distribution of cancer stage in the Waikato Breast Cancer Register

(WBCR) compared with staged cancers in the New Zealand Cancer Registry (NZCR) 1999-2011.

Table 8 shows a comparison of factors associated with stage known and unknown cancers in

the NZCR. Advanced stage, higher comorbidity score, not undergoing therapeutic surgery and

overall mortality were significantly higher for unstaged cancers (p<0.001). No significant

difference in the rate of unstaged cancer between Māori and NZ European (the two main

ethnic groups included) were observed. However, because a higher proportion of Māori

women were noted to have metastatic breast cancer compared to NZ European women (11.6%

vs. 4.7%), a separate analysis was performed for unstaged metastatic cancer by ethnicity. Of

the Māori women with unstaged cancer, 27.7% (13 out of 48) had metastatic cancer compared

to 7.8% (20 out of 257) for NZ European women, a difference which was statistically

significant (P<0.001). Overall underestimation of the incidence of metastatic breast cancer in

the NZCR was 21% (5.9% in the WBCR vs. 3.8% in the NZCR); 29.5% for Māori and 20.6%

for NZ European women, respectively.

A multivariate logistic regression was performed with unstaged cancer as the outcome variable

and age category, ethnicity, deprivation, comorbidity index and therapeutic surgery as

covariates. This identified advancing age (OR=1.63, 1.37-1.93), higher comorbidity score (OR

=1.51, 1.20-1.71) and not undergoing therapeutic surgery (OR=7.43, 5.37-10.3) as factors

significantly associated with unknown cancer stage in the NZCR (Appendix 1).

0%

10%

20%

30%

40%

50%

60%

70%

1999 2001 2003 2005 2007 2009 2011

Year of diagnosis

NZCR - Localized

WBCR - Localized

NZCR - Regional

WBCR - Regional

NZCR - Metastatic

WBCR - Metastatic


98

Table 8: Distribution of characteristics associated with stage known and unknown breast cancers in the

New Zealand Cancer Registry for the Waikato region 1999-2011.

Characteristic Total (N=2563) Stage known Stage unknown p

n % n % n %

Age group

<40 134 5.2% 124 92.5% 10 7.5% <0.001

40-59 1150 44.9% 1055 91.7% 95 8.3%

60-79 977 38.1% 877 89.8% 100 10.2%

80+ 302 11.8% 191 63.2% 111 36.8%

Ethnicity

NZ European 2077 81.0% 1820 87.6% 257 12.4%

Māori 380 14.8% 332 87.4% 48 12.6% 0.887

Pacific 50 2.0% 40 80.0% 10 20.0% 0.164

Other 56 2.2% 55 98.2% 1 1.8% 0.028

Deprivation

1-2 255 9.9% 229 89.8% 26 10.2% 0.586

3-4 257 10.0% 225 87.5% 32 12.5%

5-6 573 22.4% 504 88.0% 69 12.0%

7-8 813 31.7% 700 86.1% 113 13.9%

9-10 665 25.9% 589 88.6% 76 11.4%

Charlson Score

0 2066 80.6% 1875 90.8% 191 9.2% <0.001

1-2 418 16.3% 318 76.1% 100 23.9%

3+ 79 3.1% 54 68.4% 25 31.6%

Therapeutic Surgery

Yes 2348 91.6% 2143 91.3% 205 8.7% <0.001

No 215 8.4% 104 48.4% 111 51.6%

Outcome

Non death 1875 73.2% 1729 91.0% 170 9.0% <0.001

Death 686 26.8% 516 77.9% 146 22.1%

Deaths (n=686)

Breast cancer 409 59.6% 334 81.7% 75 18.3% <0.001

Other cause 272 39.7% 177 65.1% 95 34.9%

Unknown 5 0.7% 5 100.0% 0 0.0%

Results

99

A gradual and a significant reduction (p<0.001) in unstaged cancers and a complementary

increase in accurately staged cancers in the NZCR were observed (Figure 17). Reduction in

unstaged cancer was more pronounced from 1999 to 2004 and since had only a minimal

change. Even in 2011, approximately 12% of breast cancers included in the NZCR were either

unstaged or staged inaccurately (i.e. unstaged 6.9% and inaccurately staged 5.7%) compared

with the WBCR.

A survival analysis was performed to compare overall crude survival rate by extent of cancer

in the NZCR and the WBCR (Figure 18). Unstaged cancer in the NZCR showed a 5-year

survival of 55.9% which was between the survival rates for regional (locally invasive and/or

regional lymph node positive) at 77.3% and metastatic disease at 12.9% respectively. For

women with regional disease, both the NZCR and the WBCR exhibited almost similar

survival rates (5-year survival 77.3% vs. 76.4%). For localized disease WBCR women had a

worse survival (5-year survival 86.6% vs. 90.1%) while for metastatic cancer, the WBCR

survival was better (5-year survival 17.3% vs. 12.9%) compared with the NZCR.

Figure 17: Trends in unstaged, accurately staged and inaccurately staged breast cancer in the New

Zealand Cancer Registry compared with the Waikato Breast Cancer Register 1999-2011.

0%

20%

40%

60%

80%

100%

1999 2001 2003 2005 2007 2009 2011

Accurately staged

Unstaged

Inaccurately staged


100

Figure 18: Kaplan-Meier survival curves by cancer stage for invasive breast cancers included in the

New Zealand Cancer Registry (NZCR) and the Waikato Breast Cancer Register (WBCR) for the

Waikato region 1999-2011.

Cox proportional hazard model (Table 9) identified that unstaged cancers were associated with

a significantly higher risk of overall mortality (HR=1.59, p<0.001) compared with staged

cancers in the NZCR after adjusting for age, comorbidity index and cancer stage.

Discussion:

A high rate of overall case completeness and a high accuracy of staging for staged breast

cancer in the NZCR were observed in this study. Further, between the two registries, a high

correlation in stage distribution over the study period and roughly comparable overall survival

rates by stage was observed, despite the substantial proportion of unstaged invasive breast

cancers included in the NZCR. Although the proportion of unstaged cancer has improved, in

2011 the proportion of unknown and inaccurate staging was still 12%. Women with advanced

stage cancers, higher comorbidity score and women who were not receiving therapeutic

surgical interventions were significantly over-represented among unstaged cancers. Māori

women were significantly over-represented in unstaged metastatic breast cancer. Comparable

rates of case completeness, staging accuracy and unstaged tumours have previously been

reported for breast cancer from population based cancer registries in the United States, the

United Kingdom, the Netherlands, Denmark and Germany (72, 280-284). To our knowledge

this is the first independent audit of the breast cancer records of the NZCR performed since

mandatory reporting was introduced in New Zealand in 1994.

Results

101

Table 9: Multivariable Cox proportional hazard model for overall mortality risk for unstaged vs.

staged cancer in the New Zealand Cancer Register

Characteristic HR 95% CI p

Staging status (NZCR) Staged Ref

<0.001

Unstaged 1.59 1.32-1.92

Charlson score a 0 Ref

<0.001

1-2 2.16 1.81-2.57

3+ 3.60 2.69-4.80

Stage (WBCR) b Localized Ref

<0.001

Locally invasive 2.39 1.58-3.61

Regional LN involved 1.87 1.58-2.23

Metastatic 12.2 9.77-15.3

Age category (years) <40 Ref

<0.001

40-59 0.65 0.45-0.95

60-79 0.92 0.63-1.33

80+ 2.19 1.48-3.24

(HR – Hazard ratio, CI – Confidence interval, a Charlson Comorbidity Score, b Waikato Breast Cancer

Register).

Reasons for unstaged cancer can be grouped into two categories; lack of staging and lack of

reporting. Lack of staging occurs, for example when life expectancy is limited due to severe

comorbidities or old age, due to patient refusal and in situations where necessary staging

investigations were not available locally or where a patient could not afford investigations

(285-287). Second, where the cancer stage was known to the treating physician or recorded in

clinical documents but was not reported to the cancer registry (287). Since the breast and

axilla are relatively easily accessible areas both clinically and with simple imaging, clinical

stage at least is expected to be available for most, if not all women with breast cancer. This is

confirmed by the fact that the WBCR has been able to record cancer stage for all but one

woman, based on one or more of clinical, imaging and histopathology records.

Staging of cancers by the NZCR depends on diagnostic and therapeutic information obtained

from pathology reports and other hospital records provided by reporting laboratories and

hospitals. However, it appears that there has been a relative lack of non-pathologic (i.e.

clinical and imaging) information being provided to the NZCR, which is evident by a high rate

of unstaged cancer seen among women not undergoing therapeutic surgery. This is further


102

supported by a high proportion of metastatic breast cancer (43.7%), which in most situations is

diagnosed through imaging, being under-staged or unstaged in the NZCR. This error has led to

a significant underestimation of the true incidence of metastatic breast cancer by almost 30%

for Māori and by 20% for NZ European women. This underestimation explains the reason for

unstaged cancers exhibiting a rate of survival worse than regional disease. Similar patterns of

survival for several types of unstaged cancers in the NZCR including breast, colon and lung

have been reported by Gurney and colleagues (278).

As we have observed, unstaged cancer was more likely to be associated with metastatic

disease compared to localized or lymph node involved disease. As such, statistical analyses

which exclude these unstaged cancers, or analyses that consider these data as missing at

random and apply statistical techniques such as simple multiple imputation or inverse

probability weighting will likely lead to biased estimates of the true stage distribution (71).

Although more complex methods including multiple imputation combined with either chained

equations or stage modelling have been shown to provide more accurate estimates of stage

distribution (72), these are not used widely due to their complex nature.

This study assumes that the WBCR breast cancer data were captured perfectly without errors

from all available records. The WBCR involves collecting breast cancer data from clinical

records and pathology reports and entering data into a database by trained data entry

personnel. Close supervision by two breast surgeons and a stringent quality control and audit

process is in place to maximize the completeness, quality and accuracy of the WBCR records.

All these measures we believe have helped to minimize errors in the WBCR database and

underlie the main strength of this study.

In 2010, an independent review of the NZCR recommended an increase in breadth of data

collected, particularly through collection of clinical and imaging staging information at the

time of diagnosis (clinical TNM/cTNM) to enhance accuracy of staging and to minimize

number of unstaged cancers (288). As we have reported, more than 50% of unstaged breast

cancers were from women not undergoing therapeutic surgical interventions and using cTNM

was expected to capture clinical staging data for a majority of otherwise unstaged cancers.

Based on these recommendations, a focussed pilot project is currently being trialled by the

NZCR to identify the feasibility of collecting and relaying cTNM data through Multi-

Disciplinary Meetings (MDM) to the NZCR (288). In New Zealand, the vast majority of

cancers will be managed through MDM’s once the National Tumour Standards of Service

Provision are implemented. If this system is successful, it is expected to capture accurate

Results

103

cTNM staging data for a majority of cancers. As more structured and reliable information is

expected to be provided through synoptic reporting, the NZCR is considering a system for

automated electronic transfer of pathology information for more efficient transfer of pathology

data to the NZCR.

Population based national cancer registries including the NZCR has the objective of providing

key cancer variables such as incidence, mortality, inequities, stage and basic cancer

characteristics with a complete nationwide coverage. The NZCR has performed a

commendable job over time to provide these key cancer variables with a very high coverage,

which is on par with the top national cancer registries in the world. Despite some deficiency in

stage coverage as observed in this study, the NZCR has captured proportional as well as trends

in stage distribution over the study period with a fairly high accuracy. From an

epidemiological point of view, this evidence confirms the NZCR stage as a valid marker for

most population statistical purposes, despite some limitations in areas including metastatic

breast cancer.

The NZCR does not possess details of other important cancer related data such as diagnostic

process, treatment details and timeliness of treatment and outcomes including local and

metastatic recurrence as these aspects are beyond the scope of a national cancer registry (288).

Tumour specific regional or national registries like the regional breast cancer registries are

equipped to capture comprehensive and accurate tumour specific information. Detailed

information helps to identify quality of care issues around and to recognize where quality

improvement could be undertaken to achieve better patient outcomes. Further, there is

potential for these regional registries to be linked electronically to the NZCR in the future to

enhance accuracy and completion of NZCR data. Currently, the four regional breast cancer

registries, prospectively collect comprehensive breast cancer data from diagnosis through

treatment, follow up and outcomes. Unfortunately, lack of recognition of the importance of

these breast cancer registries and hence lack of funding is threatening the continuation of the

registries and has prevented further expansion to incorporate other regions of the country.

In conclusion, while acknowledging the commendable performance of the NZCR, we

emphasize that to increase the usefulness of the NZCR, improvements need to be made in

completeness and accuracy of staging data and rate of TNM recording. To this end, it is

crucial that all avenues for relaying cancer information to the NZCR are explored and that

appropriate methods are implemented. Improvements to the completeness and quality of data

on the NZCR will allow a more reliable estimation of important cancer issues, especially for


104

metastatic breast cancer incidence. We found an almost 30% underestimation of metastatic

breast cancer incidence for Māori compared with an almost 20% underestimation for NZ

European women. These findings provide reference for analysis of the NZCR data, in

particular for consideration of analysis of unstaged cancers by ethnicity. Alongside these

improvements in the NZCR, national and regional cancer registries need to be supported to

continue and improve to provide detailed cancer data to inform cancer control for the New

Zealand population.

Results

105

5.2. What risk factors contribute to ethnic inequities in breast cancer? A

preliminary analysis

Preface:

This chapter contains an abbreviated version of a manuscript published in Public Health

Authors: Seneviratne SA, Campbell ID, Scott N, Lawrenson R, Shirley R, Elwood M.

Title: Risk factors associated with mortality from breast cancer in Waikato, New

Zealand: A case control study

Journal: Public Health


DOI: 10.1016/j.puhe.2015.02.008

Impact factor: 1.46

Journal’s aims and scope: Public Health is an international, multidisciplinary peer-

reviewed journal. It publishes original papers, reviews and short reports on all aspects

of the science, philosophy, and practice of public health. It is aimed at all public health

practitioners and researchers and those who manage and deliver public health services

and systems. It will also be of interest to anyone involved in provision of public health

programmes, the care of populations or communities and those who contribute to

public health systems in any way.


106

Abstract:

Background:

Indigenous Māori women have a 60% higher mortality rate compared with NZ European

women. Many factors, much of which is unknown at present, are believed to be contributing to

this disparity.

Methods:

We performed a case control study to identify key characteristics associated with mortality

from breast cancer among women with newly diagnosed breast cancer between 01/01/2002

and 31/12/2010 in Waikato, New Zealand. A total of 258 breast cancer deaths were identified

from 1767 invasive cancers diagnosed over this period.

Results:

Breast cancer deaths (n=246) were compared with an age and year of diagnosis matched

control group (n=652) who were alive at the time of the death of the corresponding case and

subsequently did not die from breast cancer. Diagnosis through symptomatic presentation,

advanced stage, higher grade, absent hormone receptors (i.e. oestrogen and progesterone) and

HER-2 amplification were associated with significantly higher risks of breast cancer mortality

in bivariate analysis. Tumour stage, grade and hormone receptor status remained significant in

the multivariable model, while mode of detection and HER-2 status were non-significant. In

the bivariate analysis, Māori women had a higher risk of breast cancer mortality compared to

NZ European women (OR=1.34) which was statistically non-significant. However in the

adjusted model, risk of mortality was lower for Māori compared to NZ European women,

although this was statistically not significant (OR=0.85).

Conclusions:

Mortality pattern from breast cancer in this study were associated with established risk factors.

Ethnic inequity in breast cancer mortality in New Zealand appears to be largely attributable to

delay in diagnosis and tumour related factors. Further research in a larger cohort is needed to

identify the full impact of these factors on ethnic inequity in breast cancer mortality.

Results

107

Background

This was a preliminary study that was conducted during the early phase of the research project

with the aim of ascertaining key risk factors associated with breast cancer mortality. It was

expected that these findings would provide an insight into the factors that are to be studied and

to recognize areas needing more focus. Other objectives of this study included use of a

different study design from the cohort design which is used throughout this project. As this

study was conducted during the early phase of this research project, only women diagnosed

during 2002-2010 were included.

Methods

Study design:

A nested case-control study was performed to identify key factors associated with death from

breast cancer among women diagnosed with invasive primary breast cancer in the Waikato.

Study participants:

All women with an invasive primary breast cancer diagnosis over a 9-year period from

01/01/2002 to 31/12/2010 were identified from the WBCR (n=1919). To identify cases of

BCSM, cause of death for all deceased women from this group were identified (censored at

31/12/2012) from the Mortality Collection of the Ministry of Health and compared with same

data from the WBCR and patient clinical records.

The controls were age (+/- one year) matched women who had the same year of diagnosis and

who were alive at the time of death of the corresponding case and subsequently did not die of

breast cancer. These controls were followed up to the last recorded follow up which censored

at 31/12/2012. Controls were randomly selected and individually matched for corresponding

cases. We aimed to identify three controls per case. Same aged controls were given priority in

the selection of controls, followed by age +/- one-year controls. Age +/- one year controls

were selected randomly alternating between + and – ages based on availability. If no matched

controls were available according to above criteria, such cases were excluded from analysis.

Data analysis:

Bivariate analyses were performed using Chi squared test or Wilcoxon rank for categorical

variables and independent samples t-test for continuous variables. Odds ratios (OR) with 95%


108

confidence intervals (CI) were calculated individually and then in a multivariate model using

conditional logistic regression to identify independent risk factors associated with BCSM. All

variables with an unadjusted p value <0.20 in bivariate analyses were included in the

multivariate models. Manual stepwise backward selection procedures were used to select

variables for inclusion in multivariate models. Variables returning a p value of >0.05 were

excluded one by one, starting with the variable with the highest p value. All excluded variables

(i.e. with p>0.05) were reintroduced separately into final models and p>0.05 was reconfirmed.

Survival times were calculated from date of diagnosis to date of death in years and separate

multiple logistic regression models were developed by duration of survival of cases (≤3 or >3

years of survival).

Results

This nested case control study included 246 cases who died due to breast cancer, and 652

control women who survived (n=591) or died due to causes other than breast cancer (n=61); a

total of 898 women with invasive breast cancer. No matched controls were available for 12

cases that were mostly from extremes of age (i.e. age <30 years or >90 years) and these cases

were excluded from analyses.

Patient characteristics with bivariate analyses are summarized in Table 10. Symptomatic

presentation compared with screen detection, advanced tumour stage and grade, negative ER

and PR status and positive HER-2 status were associated with significantly (P<0.001) elevated

risks of death from breast cancer.

A multivariable analysis was performed with conditional logistic regression (Table 11).

Duration of symptoms was not included despite being p<0.001, since it was available only for

women with symptomatically detected breast cancer. The risk of death from breast cancer was

significantly (P<0.001) increased with more advanced tumour stage and grade and with

negative hormone receptor status; it was also increased, although not significantly, with

positive HER-2 status (p=0.385). A separate model with ethnicity adjusting only for cancer

stage yielded an OR of 0.92 (95% CI 0.59-1.58, p=0.89) for BCSM for Māori compared to NZ

European women.

Results

109

Table 10: Bivariate analysis of risk factors and mortality

Variable

Cases (N=246)

n (%)

Controls (N=652)

n (%) OR 95% CI p

Mean age ± SD 61.1 ± 14.6 60.6 ± 13.9 0.591

Detection method <0.001

Screen detected 35 (14.2) 250 (38.3) Ref

Non-screen 211 (85.8) 402 (61.7) 3.75 2.54 - 5.54

Mean symptom

duration ± SD a 14.9 ± 21.8 10.1 ± 14.6 0.193

Ethnicity

NZ European 192 (78.0) 540 (82.2) Ref

Māori 43 (17.5) 90 (13.8) 1.34 0.90 - 2.00 0.154

Pacific 7 (2.8) 11 (1.69) 1.79 0.68 - 4.68 0.243

Other 4 (1.6) 11 (1.7) 1.03 0.32 - 3.25 0.971

Tumour stage <0.001

I 16 (6.5) 271 (41.6) Ref

II 79 (32.1) 287 (44.0) 4.66 2.66 - 8.18

III 84 (34.1) 88 (13.5) 16.2 9.00 - 29.1

IV 67 (27.2) 6 (0.9) 189.1 71.3 - 501.7

Tumour grade <0.001

I 8 ( 4.2) 174 (29.0) Ref

II 83 (43.5) 322 (53.6) 5.61 2.65 - 11.8

III 100 (52.4) 105 (17.5) 20.7 9.69 - 44.3

Unknown (55) (51)

ER/PR <0.001

Positive 172 (72.6) 550 (89.6) Ref

Negative 65 (27.4) 64 (10.4) 3.25 2.21 - 4.78

Unknown (9) (38)

HER-2 <0.001

Negative 120 (60.9) 348 (72.8) Ref

Equivocal 15 (7.6) 59 (12.3) 0.74 0.40 - 1.35

Positive 62 (31.5) 71 (14.9) 2.53 1.77 - 3.77

Unknown (49) (174)


110

Histology

Ductal 193 (85.4) 525 ( 81.3) Ref

Lobular 23 (10.2) 73 (11.3) 0.86 0.52 - 1.41 0.583

Other 10 (4.4) 48 (7.4) 0.56 0.26 - 1.40 0.132

Unknown (20) (6)

Charlson Index 0.951

0 220 (89.4) 581 (89.1) Ref

1-2 8 (3.3) 24 (3.7) 0.88 0.39 - 1.98

3+ 18 (7.3) 47 (7.2) 1.01 0.57 - 1.78

(OR – odds ratio, CI – confidence interval, a - symptomatic cancers only)

To identify factors associated with duration of survival among women with BCSM, cases were

categorized into two groups [survival length ≤3 years (n=199) and >3 years (n=74)]. Using

individually matched corresponding controls, separate logistic regression analyses with

conditional regression were performed for these two groups (Table 12). Tumour grade showed

a stronger associations with mortality ≤3 years compared to >3 years and, ER/PR negativity

was associated with increased deaths only before 3 years, although the differences between the

two time groups were not significant. HER-2 status was associated with increased deaths only

after 3 years (OR 2.49, 95% CI 1.01-6.21) compared to HER-2 negative women (Table 12).

HER-2 receptor status was available for 73.2% (536 out of 732) NZ European and 86.5% (115

out of 133) Māori women. Of this, HER-2 amplification was seen among 92 (17.2%) NZ

European and 33 (28.7%) Māori women. Only 3.2% (2 out of 62) of HER-2 amplified cases

received treatment with trastuzumab (Herceptin®) compared with 25.4% (18 out of 71) of

HER-2 amplified controls. To identify the impact of trastuzumab on survival among HER-2

positive women, a separate regression analysis was performed inclusive of trastuzumab

treatment. This demonstrated trastuzumab treatment to be associated with a significantly lower

risk of BCSM among women with HER-2 amplified tumours (p=0.003, OR 0.07, 95% CI

0.01–0.42). Of the HER-2 amplified tumours, 14 (15.2%) NZ European and 4 (12.1%) Māori

women received treatment with trastuzumab.

Results

111

Table 11: Multivariate model for factors associated with mortality

Variable OR 95% CI p

Detection method 0.085

Screen detected Ref

Symptomatic 1.52 0.94 – 2.45

Ethnicity 0.567

NZ European Ref

Māori 0.85 0.49 – 1.48

Tumour stage <0.001

I Ref

II 2.68 1.47 – 4.90

III 9.08 4.78 – 17.3

IV 111.1 35.8 – 344.8

Tumour grade <0.001

I Ref

II 2.98 1.31 – 6.81

III 6.88 2.92 – 16.2

ER/PR 0.001

Positive Ref

Negative 2.43 1.49 – 3.97

HER-2 0.385

Negative Ref

Equivocal 1.32 0.63-2.79

Positive 1.54 0.92-2.60

(OR – odds ratio, CI - confidence interval)


112

Table 12: Multivariate model (conditional logistic regression) for factors associated with mortality for

women with a survival of ≤3 years and >3 years against matched controls

Variable

Case survival ≤3 years (n=199) Case survival >3 years (n=74)

OR 95% CI p OR 95% CI p

Detection method 0.013 0.967

Screen detected Ref Ref

Symptomatic 2.41 1.21-4.81 0.98 0.48-2.00

Ethnicity 0.835 0.402

NZ European Ref Ref

Māori 1.08 0.54-2.16 0.64 0.22-1.82

Tumour stage <0.001 <0.001

I Ref Ref

II 2.67 1.15-6.19 2.80 1.14-6.87

III 10.2 4.29-24.3 7.05 2.59-19.2

IV 122.6 30.3-496.1 107.6 10.5-980.1

Tumour grade <0.001 0.151

I Ref Ref

II 8.61 1.63-45.4 1.91 0.71-5.15

III 23.9 4.45-125.4 3.58 1.15-11.2

ER/PR 0.001 0.837

Positive Ref Ref

Negative 3.17 1.76-5.72 1.37 0.49-3.82

HER-2 0.978 0.275

Negative Ref Ref

Equivocal 1.04 0.39-2.80 1.43 0.44-4.67

Positive 1.16 0.60-2.24 2.49 1.01-6.21

(OR – odds ratio, CI - confidence interval)

Results

113

Discussion

In this case-control study we observed a pattern in concordance with known risk factors for

breast cancer specific mortality (BCSM); i.e., tumour stage, grade and hormone receptor

(ER/PR) status were independently associated with BCSM.

In the univariate analysis, Māori women had an estimated 34% higher risk of breast cancer

death; while not significant, this is compatible with previous analyses which have assessed

breast cancer mortality in Māori women (4, 31). For example, a study based on the NZCR data

from 2002 to 2006 has shown a 73% higher age standardized breast cancer mortality rate for

Māori compared to non-Māori women (4). However, no increased risk of death in Māori

women was seen after taking into account the major predictors of stage, grade, and hormone

receptor status, the adjusted odds ratio being 0.85. Even adjusting only for stage completely

negated the increased risk of breast cancer death that was observed among Māori in the

univariate analysis. This indicates that the increased death risk for Māori women with breast

cancer may be largely due to presentation at a more advanced stage, similar to findings of

Jefferys et al (9), suggesting that further efforts to improve early diagnosis are needed.

Reporting of HER-2 receptor status has only been routine in New Zealand since 2007. One

recently published study based on the NZCR data has reported a significantly higher rate of

HER-2 amplified tumours in Māori compared to non-Māori non-Pacific women (18.4% vs.

25.7%) (13). Trastuzumab (Herceptin) which has shown to improve survival among women

with HER-2 amplified breast cancer (289) became available for the treatment of HER-2

amplified early breast cancer through the publicly funded system in New Zealand from 2007

(61). Prior to 2007, trastuzumab was available only for women who could privately fund this

expensive therapy. A substantially higher proportion of HER-2 amplified breast cancer and a

marginally lower rate of trastuzumab therapy for HER-2 amplified tumours were seen in

Māori compared to NZ European women in our study. However, the significance of this

treatment difference is uncertain due to small number of women receiving trastuzumab

treatment included in this study. Ethnic inequities are known to exist even for publicly funded

medications in New Zealand. For example in 2000, Māori and Pacific patients were 60% less

likely to receive statin therapy despite having higher incidences of cardiovascular disease

compared to Europeans (290). This raises an issue of ensuring prompt access to new therapies

for Māori women particularly when they are more likely to have a greater benefit from such

treatment.


114

This study has some limitations. The follow up duration was relatively short, with a median

follow up of 6 years. As mortality from breast cancer is known to occur even after 20 years,

this study is more likely to highlight the factors associated with short and medium term

BCSM. The limited sample of Māori cases included in this study prevented a separate logistic

regression analysis being performed for Māori women, which could have provided us with

more information on the differences in distribution of risk factors and their impact on observed

mortality disparity between Māori and European ethnic groups. We plan to examine these

factors further in a larger cohort.

In conclusion, this is the first study of its kind in New Zealand investigating the association

between death from breast cancer and, patient and tumour biological features from a

comprehensive data set. These findings have important implications for targeting early

diagnosis and providing access to targeted therapies especially for Māori women which could

potentially reduce ethnic inequities in breast cancer mortality in New Zealand. Further

research with a larger cohort of women is needed and is currently underway to further assess

the impact of patient, tumour and treatment characteristics on ethnic inequity in breast cancer

mortality in New Zealand.

Results

115

5.3. Is breast cancer screening contributing to ethnic inequity in outcomes?

Preface:

This chapter contains an abbreviated version of a manuscript published in BMC Public Health

Authors: Seneviratne S, Campbell I, Scott N, Shirley R, Lawrenson R.

Title: Impact of mammographic screening on ethnic and socioeconomic inequities in

breast cancer stage at diagnosis and survival in New Zealand: A cohort study

Journal: BMC Public Health


DOI: 10.1186/s12889-015-1383-4

Impact factor: 2.32

Journal’s aims and scope: BMC Public Health is an open access, peer-reviewed

journal that considers articles on the epidemiology of disease and the understanding of

all aspects of public health. The journal has a special focus on the social determinants

of health, the environmental, behavioural, and occupational correlates of health and

disease, and the impact of health policies, practices and interventions on the

community.


116

Abstract:

Background:

Breast cancer screening coverage for Indigenous Māori women is significantly lower

compared with NZ European women, and is believed to be contributing to more advanced

cancer at diagnosis and worse survival in Māori women. This study was done to explore the

impact of differences in rates of screen detected breast cancer on inequities in cancer stage at

diagnosis and survival between Māori and NZ European women in the Waikato.

Methods:

All primary breast cancers diagnosed in screening age women (as defined by the BSA

Programme) during 1999-2012 in the Waikato area (n=1846) were identified from the WBCR

and the National Screening Database. Stage at diagnosis and survival were compared for

screen detected (n=1064) and non-screen detected (n=782) breast cancer by ethnicity and

socioeconomic status.

Results:

Indigenous Māori women were significantly more likely to be diagnosed with more advanced

cancer compared with NZ European women (OR=1.51), and approximately a half of this

difference was explained by lower rate of screen detected cancer for Māori women. For non-

screen detected cancer, Māori had significantly lower 10-year breast cancer survival compared

with NZ European (46.5% vs. 73.2%) as did most deprived compared with most affluent

socioeconomic quintiles (64.8% vs. 81.1%). No significant survival differences were observed

for screen detected cancer by ethnicity or socioeconomic deprivation.

Conclusions:

The lower rate of screen detected breast cancer appears to be a key contributor towards the

higher rate of advanced cancer at diagnosis and lower breast cancer survival for Māori

compared with NZ European women. Among women with screen-detected breast cancer,

Māori women do just as well as NZ European women, demonstrating the success of breast

screening for Māori women who are able to access screening. Increasing breast cancer

screening rates has the potential to improve survival for Māori women and reduce breast

cancer survival inequity between Māori and NZ European women.

Results

117

Background:

In parallel with many other developed countries, New Zealand has experienced a substantial

reduction in breast cancer mortality over the last two decades (1). While some of the observed

reduction in breast cancer mortality in these countries has been attributed to advances in

treatment, mammographic breast cancer screening has also played a major role (130). Since it

was established in 1999, the BSA has provided free biennial mammographic screening for all

women aged between 50 to 64 years and this age range was extended to include women aged

45 to 49 and 65 to 69 years in July 2004.

Mammographic screening coverage in New Zealand has gradually picked up over the last

decade and has achieved the target biennial coverage of 70% for NZ European women since

2010 (39). However, poor screening coverage has remained a significant issue for Māori

women for whom the coverage was only 62.7% in 2012 (39), well below the 70% target

coverage. Further, there was a large variability in screening coverage rate for Māori by region,

which ranged from 54% to 79% across the country in 2012 (40). There is, however, close

monitoring of Māori coverage, and targets are set yearly with screening providers to improve

this. It is this effort that has picked Māori coverage up from below 40% at the commencement

of BSA programme in 1999 to the current level of over 60% (40). In spite of this, lower

screening coverage remains a likely major contributor to inequities in breast cancer burden

between Māori and NZ European women, as Māori have a higher breast cancer incidence and

are more likely to present with advanced cancer compared with NZ European women (1).

This study was conducted to explore differences in rates of screen detected cancer by ethnicity

and socioeconomic deprivation in a cohort of screening age women, and to determine the

contribution of these differences to ethnic and socioeconomic inequities in breast cancer

survival in New Zealand.

Methods:

Study population:

All screening age women with newly diagnosed breast cancer between 01/01/1999 and

31/12/2012 identified from the WBCR were eligible for this study (n=1846). Screening age

was defined according to the BSA. This included women between 50 and 64 years up to June

2004 and women between 45 and 69 years from July 2005 onwards. Screening status was


118

classified into screen-detected (n=1064, 57.6%), interval (if diagnosis within 24 months from

last screening mammogram, n=241, 13.1%) and non-interval symptomatic (n=541, 29.3%).

Cancers diagnosed through BSA (n=954, 89.7%) and through opportunistic screening

mammograms arranged by physicians outside BSA (n=110, 10.3%) were included under

screen detected cancers. Screening status for each woman diagnosed through BSA was

confirmed by comparing with screening data from the BSA database, which include data (i.e.,

screen detected and interval) for all women with breast cancer diagnosed through the BSA

programme. Details of opportunistic screening were based on the WBCR records, and were

reconfirmed by accessing clinical and mammographic records of all these women. For women

with more than one episode of breast cancer during the study period, only the first cancer was

included for analysis.

Outcome variables:

Date and cause of death for all deceased women (censored at 31/12/2013) were identified from

the WBCR and from the Mortality Collection of the Ministry of Health. Follow up duration

was calculated from the date of diagnosis to the date of death, or to the date of the last follow

up when the patient was known to be alive (censored at 31/12/2013).


Chi squared (χ²) test for trend was used to test for univariate differences between groups of

interest. Multivariate logistic regression analyses were used to test for differences in early

versus advanced stage at diagnosis between Māori and NZ European women sequentially

adjusting for age, year of diagnosis, screening status, socioeconomic deprivation and

residential status. Impact of differences in screen detected cancer towards mortality disparity

between Māori and NZ European women was explored in Cox proportional hazard models

sequentially adjusting for same covariates. Interaction terms were included into regression

models to identify possible interactions between ethnicity, deprivation and screening status

(i.e., ethnicity x deprivation, ethnicity x screening and deprivation x screening). We also

investigated 5-year and 10-year breast cancer specific survival rates for invasive cancers by

screening status, ethnicity and socioeconomic deprivation using Kaplan-Meier survival curves.

Survival comparisons by ethnicity were performed for Māori and NZ European women.

Pacific and Other ethnic group women were excluded from these analyses.

Results

119

Results:

This study included a total of 1846 screening age women (1548 invasive and 298 in-situ) with

newly diagnosed first primary breast cancer over the study period. Of these women 1064

(57.6%) were screen detected and 782 (42.4%) were symptomatic. Of symptomatic women,

241 (30.8%) had interval and 541 (69.2%) had non-interval symptomatic cancers. Mean

follow-up was 65.5 months (median 58 months). Sixty-seven percent of women were followed

up for a minimum of five years or until death and 32% were followed up for ten years or until

death. Overall mean age of included women was 56.8 years (median 56). Mean ages for in-situ

and invasive breast cancers were 56.5 (median 56.5) and 56.9 years (median 56) respectively.

Lower proportions of cancers were observed within 45-49 and 65-69 age categories as these

two age categories were included in the BSA only from July 2004 onwards (Table 1). When

women diagnosed since July 2004 were considered alone, almost similar proportions of

women were observed to be included in each age category, which ranged from 19.0% (55-59

age category) to 21.1% (60-64 age category) .

No difference in age distribution was seen for Māori and NZ European women where the

median age at diagnosis was 56 years. Women of higher socioeconomic deprivation quintiles

were significantly older at diagnosis compared with women of lower deprivation quintiles

which ranged from a mean age of 55.2 years for deprivation quintile 1-2 to a mean age of 57.2

years for deprivation quintile 9-10 (p<0.001).

Overall, 62.7% of all breast cancers among NZ European women within screening age were

screen detected compared with 49.2% among Māori women of this age range (p<0.001). Of

the screen detected cancers, a higher proportion of cancers in NZ European women were

detected through opportunistic screening (10.9%) compared with Māori (7.3%), but this

difference was statistically not significant (p=0.183). The difference in proportion of screen

detected cancer between NZ European and Māori women was significant for invasive cancers

(58.3% vs. 45.1%, p=<0.001), but was not statistically significant for in-situ cancers (85.4%

vs. 74.4%, p=0.063).


120

Table 13: Characteristics of women associated with early stage [compared with advanced stage] at diagnosis of breast cancer by ethnicity for screening age women with

newly diagnosed breast cancer in the Waikato, New Zealand 1999-2012

NZ European Māori All cancers

Characteristic (N=1459) Early a (N=311) Early a (N=1846) Early a

n (%) n (%) n (%) n (%) n (%) n (%) OR 95% CI p

Screening status <0.001

Screen detected 915 (62.7) 654 (71.5) 153 (49.2) 107 (69.9) 1106 (59.9) 786 (71.1) Ref

Interval 192 (13.2) 63 (32.8) 20 (6.4) 3 (15.0) 218 (11.3) 69 (31.7) 5.30 4.15-6.74

Non-screen 352 (24.1) 118 (33.5) 138 (44.4) 36 (26.1) 522 (28.3) 162 (31.0) 5.46 4.11-7.33

Age group (years)

45-49 208 (14.3) 110 (52.9) 51 (16.4) 23 (45.1) 275 (14.9) b 143 (52.0) Ref 0.005

50-54 375 (25.7) 206 (54.9) 77 (24.8) 31 (40.3) 469 (25.4) 245 (52.2) 0.99 0.74-1.33

55-59 323 (22.1) 180 (55.7) 74 (23.8) 32 (43.2) 416 (22.5) 218 (52.4) 0.98 0.73-1.33

60-64 337 (23.1) 196 (58.2) 73 (23.5) 36 (49.3) 425 (23.0) 241 (56.7) 0.83 0.61-1.12

65-69 216 (14.8) 143 (66.2) 36 (11.6) 24 (66.7) 261 (14.1) b 170 (65.1) 0.58 0.41-0.82

Results

121

Deprivation

Dep 1-2 184 (12.6) 101 (54.9) 9 (2.9) 1 (11.1) 203 (11.0) 108 (53.2) Ref 0.094

Dep 3-4 157 (10.8) 104 (66.2) 23 (7.4) 13 (56.5) 186 (10.1) 120 (64.5) 0.63 0.42–0.94

Dep 5-6 339 (23.2) 189 (55.8) 62 (19.9) 29 (46.8) 411 (22.3) 223 (54.3) 0.96 0.68–1.34

Dep 7-8 492 (33.7) 279 (56.7) 95 (30.5) 43 (45.3) 614 (33.3) 338 (55.0) 0.93 0.68–1.28

Dep 9-10 287 (19.7) 162 (56.4) 122 (39.2) 60 (49.2) 432 (23.4) 228 (52.8) 1.02 0.73–1.42

Residence

Urban 783 (53.7) 435 (55.6) 150 (48.2) 75 (50.0) 990 (53.6) 539 (54.4) Ref 0.794

Semi-urban 370 (25.4) 222 (60.0) 107 (34.4) 50 (46.7) 492 (26.7) 277 (56.3) 0.93 0.75–1.15

Rural 306 (21.0) 178 (58.2) 54 (17.4) 21 (38.9) 364 (19.7) 201 (55.2) 0.97 0.76–1.23

Year of diagnosis

1999-2002 249 (17.1) 128 (51.4) 44 (14.1) 20 (45.5) 300 (16.3) 153 (51.0) Ref 0.423

2003-2006 450 (30.8) 270 (60.0) 75 (24.1) 33 (44.0) 542 (29.4) 308 (56.8) 0.79 0.60-1.05

2007-2009 363 (24.9) 203 (55.9) 86 (27.7) 45 (52.3) 479 (25.9) 263 (54.9) 0.86 0.64-1.14

2010-2012 397 (27.2) 234 (58.9) 106 (34.1) 48 (45.3) 525 (28.4) 293 (55.8) 0.82 0.62-1.09

(a early stage = in-situ & stage I, b only cancers diagnosed from July 2004 onwards are included)


122

Univariate analysis of factors associated with early stage disease (in-situ & stage I, n=1017) at

diagnosis compared with more advanced stage (stages II, III & IV, n=829) is shown in Table

13. Non-screen compared with screen detection was significantly associated with more

advanced stage at diagnosis (OR=5.46, p<0.001) as did Māori compared with NZ European

ethnicity (OR=1.51, 95% CI 1.18-1.93, p=0.001). No significant differences were observed in

early versus advanced stage at diagnosis by deprivation status (p=0.094). Table 14 shows odds

ratios from multivariate logistic regression analyses for advanced versus early stage at

diagnosis in Māori compared with NZ European women with sequential adjustment for

covariates. Adjusting for screening status reduced the odds ratio for more advanced stage at

diagnosis in Māori compared with NZ European from 1.49 (1.15-1.91) to 1.25 (0.96-1.64). A

minimal further attenuation in odds ratio was observed with additional adjustments for

socioeconomic deprivation and residential status (OR=1.24, 0.95-1.65). Similarly, odds of

advanced stage at diagnosis remained largely unchanged after introducing interaction terms

into the model (OR=1.28, 0.74-2.23). In a separate analysis (data not shown), socioeconomic

deprivation was introduced into the model prior to screening status, but made no difference to

the odds ratio (OR=1.49, 1.16-1.93).

Table 14: Odds ratios for stage at diagnosis (i.e., advanced b versus early a) in Māori compared with

NZ European women with stepwise adjustment for age, year of diagnosis, screening status,

socioeconomic deprivation and urban/rural residential status

Characteristic OR 95% CI p

Model A (Unadjusted)) 1.51 1.18–1.93 0.001

Model B (Age adjusted) 1.49 1.16-1.91 0.002

Model C (Model B + Year of diagnosis c) 1.49 1.15-1.91 0.002

Model D (Model C + Screening status) 1.25 0.96-1.64 0.101

Model E (Model D + Deprivation) 1.24 0.94-1.64 0.133

Model F (Model E + Urban/Rural residence) 1.24 0.95-1.65 0.125

Model H (Model E + interaction terms d) 1.28 0.74-2.23 0.373

(a early stage = in-situ & stage I, b advanced stage = stages I to III, c year categories as in Table

1, d – ethnicity*deprivation, ethnicity*screening and deprivation*screening)

Results

123

Next we repeated a similar regression analyses only including invasive cancers to identify

factors associated with more advanced invasive cancer (stages II, III & IV, n=829) versus

early invasive (stage I, n=719) cancer at diagnosis (data not shown). This analysis yielded

results similar to initial analysis with Māori having significantly more advanced disease at

diagnosis (unadjusted) compared with NZ European women (OR=1.53, 1.17-2.01, p=0.002).

Adjusting for screening status resulted in a reduction of age and year adjusted odds ratio for

advanced cancer in Māori compared with NZ European from 1.52 (1.15-1.99) to 1.30 (0.97-

1.75). Further adjustment for deprivation made no difference to this odds ratio (OR=1.30,

0.96-1.75).

Table 15: Adjusted (age and year of diagnosis) breast cancer specific mortality hazard ratios from Cox

regression model

All cancers Screen detected Non-screen detected

Characteristic HR 95% CI p HR 95% CI HR 95% CI p

Ethnicity

NZ European Ref Ref Ref

Māori 1.33 0.33-3.18 0.964 1.45 0.33-6.39 0.620 3.13 1.58-6.18 0.001

Mode of diagnosis

Screen detected Ref - -

Non-Screen 2.81 1.57-5.04 0.001

Deprivation quintile

1-2 Ref 0.118 Ref 0.627 Ref 0.019

3-4 0.84 0.39-1.78 0.77 0.19-3.10 0.76 0.35-2.13

5-6 0.86 0.45-1.66 0.91 0.28-2.79 0.68 0.39-1.84

7-8 1.26 0.58-2.74 1.13 0.35-3.60 0.08 0.93-3.76

9-10 0.73 0.32-1.69 0.56 0.15-2.06 0.71 0.52-2.56

Residential status

Urban Ref 0.311 Ref 0.370 Ref 0.155

Semi-urban 0.77 0.53-1.12 1.51 0.73-3.14 0.65 0.42-1.01

Rural 0.81 0.54-1.21 0.78 0.31-1.94 0.85 0.54-1.32

Ethnicity x Deprivation 0.65 0.31-1.36 0.253 0.39 0.05-3.07 0.370 0.69 0.31-1.54 0.468

Ethnicity x Screening 3.29 1.10-9.85 0.033 - -

Deprivation x Screening 1.47 0.71-3.06 0.302 - -


124

Adjusted breast cancer specific mortality hazard ratios from Cox regression models by

screening status for women with invasive breast cancers are shown in Table 15. Māori women

were observed to have higher hazards of breast cancer mortality overall (HR=1.33, 0.33-3.18)

and for screen detected (HR=1.45, 0.33-6.39) and non-screen detected cancers (HR=3.13,

1.58-6.18), although this was statistically significant only for non-screen detected cancer.

Breast cancer specific mortality hazard ratios for Māori compared with NZ European women

with sequential adjustment for covariates is shown in Table 16. In the model for all cancers,

adjusting for screening status reduced the mortality hazard for Māori from 2.33 to 2.01, while

further adjusting for deprivation resulted in a marginal increase of hazard up to 2.09.

Adjusting for socioeconomic deprivation before adjusting for screening (data not shown) saw

a similar increase in hazard for Māori from 2.33 to 2.37. For screen detected cancers, Māori

women had non-significant lower hazards of mortality before (HR=0.77) and after adjusting

for covariates (HR=0.85), but was non-significantly higher after introduction of interaction

terms (HR=1.45). In contrast, for non-screen detected cancers Māori had a significantly higher

risk of mortality which was more than double that for NZ European women before (HR=2.28)

and after adjusting for covariates (HR=2.37) and introduction of interaction terms (HR=3.13).

Unadjusted five and 10-year breast cancer specific survival rates for invasive cancers by

screening status, ethnicity and deprivation are shown in Table 17. Screen detected invasive

cancers demonstrated the highest five and 10-year breast cancer specific survival rates and the

lowest rates were seen for non-interval symptomatic breast cancers. Māori women had

significantly lower crude five (90.2% vs. 77.6%) and 10-year (83.5% vs. 73.8%) breast cancer

survival rates compared with NZ European women. Compared with women from more

affluent (1-2, 3-4 to 5-6) quintiles, women from more deprived quintiles (7-8 and 9-10) had

significantly worse 10-year breast cancer survival rates (84.3% vs. 77.2%, p=0.011).

Māori women with non-screen detected invasive cancer had significantly worse five and 10-

year breast cancer survival rates compared with NZ European women (83.1% vs. 64.2% and

73.2% vs. 46.5% respectively, p<0.001). In contrast, for screen detected breast cancer, Māori

women had, if anything, better five and 10-year breast cancer specific survival rates (96.9%

and 94.1%) compared with NZ European women (95.8% and 90.3%), although this was

statistically non-significant (p=0.651) (Figure 19).

Results

125

Table 16: Hazard ratios for breast cancer-specific mortality risk in Māori compared with NZ European women with stepwise adjustment for age, year of diagnosis,

screening status, socioeconomic deprivation and urban/rural residential status

Characteristic

All cancers Screen detected Non-screen detected

HR (95% CI) p HR (95% CI) p HR (95% CI) p

Model A (Unadjusted)) 2.25 (1.62-3.12) <0.001 0.77 (0.27-2.15) 0.617 2.28 (1.59-3.26) <0.001

Model B (Age adjusted) 2.29 (1.69-3.18) <0.001 0.80 (0.29-2.25) 0.674 2.34 (1.63-3.35) <0.001

Model C (Model B + Year of diagnosis a) 2.33 (1.64-3.25) <0.001 0.84 (0.31-2.36) 0.738 2.27 (1.59-3.25) <0.001

Model D (Model C + Screening status) 2.01 (1.44-2.80) <0.001 - - - -

Model E (Model D + Deprivation) 2.09 (1.49-2.94) <0.001 0.85 (0.30-2.40) 0.762 2.39 (1.65-3.46) <0.001

Model F (Model E + Urban/Rural residence) 2.11 (1.50-2.97) <0.001 0.85 (0.30-2.41) 0.760 2.37 (1.64-3.47) <0.001

Model G (Model F + Interaction terms b) 1.33 (0.33-3.18) 0.964 1.45 (0.33-6.39) 0.620 3.13 (1.58-6.18) 0.001

(a year categories as in Table 1, b – ethnicity x deprivation, ethnicity x screening and deprivation x screening


126

Table 17: Five-year and 10-year breast cancer specific survival rates by screening status, ethnicity and

socioeconomic deprivation for screening age women with invasive breast cancer in the Waikato, New

Zealand 1999-2012

Characteristic No. of

women

Total breast

cancer deaths

5-year

survival 95% CI

10-year

survival 95% CI

Screening status

Screen detected 858 40 96.2% 94.6 - 97.8 91.8% 88.3 - 95.3

Interval 217 36 83.9% 79.3 - 88.5 73.5% 64.1 - 82.9

Non-screen non-interval 473 108 76.4% 70.9 - 81.9 66.3% 60.2 - 72.4

Ethnicity

NZ European 1220 126 90.2% 88.2 - 92.2 83.5% 77.4 - 89.6

Māori 268 50 77.6% 71.5 - 83.7 67.8% 58.4 - 77.2

Deprivation

Dep 1-2 165 14 89.5 83.8 - 95.2 87.1 80.2 - 94.0

Dep 3-4 155 15 89.7 84.2 - 95.2 85.8 79.7 - 93.9

Dep 5-6 347 33 90.8 87.1 - 94.5 85.0 76.6 - 89.6

Dep 7-8 513 81 84.1 80.4 - 87.8 75.4 69.9 - 80.9

Dep 9-10 368 41 88.2 84.7 - 91.7 79.7 72.6 - 86.8

Figure 19: Kaplan-Meier curves for breast cancer specific survival for screen detected (Panel A) and

symptomatically detected (Panel B) breast cancers in screening age women by ethnicity

Panel A

Log rank test p= 0.651

Panel B

Log rank test p= <0.001

Results

127

Figure 20 shows the association between screening status and socioeconomic deprivation on

10-year breast cancer specific survival rates based on Kaplan-Meier survival analysis. There

was a tendency for higher survival in women from more affluent quintiles. The difference in

breast cancer specific survival rates between more and less affluent women was substantially

lower for women with screen detected breast cancer than for non-screen detected breast

cancer. This has resulted in a greater survival difference between screen and non-screen

detected cancer with increasing socioeconomic deprivation.

Figure 20: Ten-year breast cancer specific survival rates by socioeconomic deprivation (based on

Kaplan-Meier survival curves by socioeconomic deprivation quintile) for screening age women

Figure 21 illustrates breast cancer specific survivals for women with non-screen detected

cancer by socioeconomic deprivation (Dep 1-6 versus Dep 7-10) and ethnicity. Among NZ

Europeans, women of higher socioeconomic groups (Dep 1-6) had significantly better survival

rates compared with women of lower socioeconomic groups (Dep 7-10) (10-year survival 81%

vs. 67.6%, p=0.026). In comparison, survival rates did not vary by deprivation status among

Māori women with non-screen detected cancer (10-year survival 51.2% vs. 44.4%, p=0.715).

86.8%

79.7%

91.8% 91.5%

81.1%

64.8%

0%

20%

40%

60%

80%

100%

Dep 1-2 Dep 3-4 Dep 5-6 Dep 7-8 Dep 9-10

All invasive cancers

Screen detected

Non-screen detected


128

NZ European Māori

Figure 21: Kaplan-Meier survival curves for non-screen detected cancers in screening age NZ

European and Māori women by socioeconomic deprivation status

Discussion:

From this study we found that breast cancer survival inequities between Māori and NZ

European women in the screening age group were significant for women diagnosed through

the non-screen pathway, but absent for women diagnosed through screening. Five and 10-year

survival for women diagnosed through screening were 2.8% and 5.2% higher respectively, for

Māori compared with NZ European women, while Māori women diagnosed through the non-

screen pathway had a 9.9% lower 5-year and a 17.7% lower 10-year survival than NZ

European women. We found that patterns of inequities by socioeconomic deprivation were

similar to those found between Māori and NZ European women. For non-screen detected

cancer, 10-year breast cancer specific survival was 16.3% lower for most deprived compared

with women from most affluent socioeconomic quintile.

Benefits of population based mammographic breast cancer screening have been proven by

several randomized trials (130, 291). It has been shown that provision of biennial screening

with a 70% coverage confers an approximately a 30% reduction in breast cancer mortality for

the screening population (292). The breast cancer incidence for NZ European women is

similar to the rates seen in breast cancer screening trials (55, 292), but incidence in Māori is

much higher (293). Hence, at a 70% biennial screening rate, more Māori women may benefit

than non-Māori, and given that Māori women with symptomatic cancer tend to present at a

later stage than non-Māori, Māori women may also have a greater reduction in breast cancer

mortality with screening. Our data lend support to this hypothesis.

Results

129

Compared with NZ Europeans, a higher proportion of breast cancer is observed among

younger Māori women due to the younger age structure of Māori population (164). Further,

the screen detection rate for Māori women between 45 to 49 years in initial mammographic

screens is more than twice that for non-Māori, and is 50% greater even for subsequent

screening mammograms (40). Breast cancer in younger women tends to be more aggressive

and is more likely to be associated with poorer outcomes (294). Thus, it may be worthwhile

exploring the usefulness of providing breast cancer screening for Māori women below the

current screening age limit, for instance for women between 40 and 44 years. However, a

thorough evaluation of potential benefits versus harms of mammographic screening should be

an essential part of such an initiative.

Obese women are more likely to present with more advanced breast cancer compared with

non-obese women and are more likely to be diagnosed through screening mammography than

non-obese (295). Approximately 48% of adult Māori women are obese, which is almost twice

the rate observed in NZ European women (33). Evidence from the USA suggests that up to

30% of later stage breast cancer diagnosed in African American women compared with White

American women are attributable to higher rates of obesity in African American women (296).

At present, no New Zealand data are available on the impact of obesity on ethnic differences

in breast cancer stage at diagnosis or outcomes. Nonetheless, higher obesity prevalence is

likely to be another factor that would confer a greater benefit from increased mammographic

screening coverage for Māori compared with NZ European women.

Survival differences for screen detected breast cancer were not seen, either by ethnicity or

socioeconomic deprivation, while differences in survival were marked and significant for non-

screen detected cancer by both. Similar observations have been reported from the UK, the

USA and the Netherlands (137, 293, 297). Inequities in breast cancer survival by ethnicity and

socioeconomic status for women diagnosed through non-screen pathway are likely due to a

range of factors. While later stage at diagnosis for Māori and women of socioeconomically

deprived groups is the likely major factor for these differences, higher rates of obesity,

comorbidities and differences in treatment compared with NZ European and more affluent

women, respectively are also likely to be important. In addition, women who participate in

mammographic screening may be more likely to have better access to health care, education,

health literacy and health seeking behaviours compared with women from similar ethnic or

socioeconomic backgrounds, who do not participate in screening (298).


130

Furthermore, there is a stringent quality framework and an audit process for the treatment of

breast cancer detected through BSA, but not for cancers diagnosed outside BSA (39). Results

for each regional screening provider are regularly audited and a feedback process ensures that

adequate measures are undertaken by providers who fail to achieve these targets. Similar audit

processes or quality frameworks are not in place for non-screen detected cancer treatment.

Lack of a framework is a likely key reason for the disparities observed amongst women with

non-screen detected cancers by ethnic and socioeconomic grouping. To overcome these

disparities in cancer care, the Ministry of Health is in the process of introducing quality

measures for the management of all common cancers through the National Cancer Control

Strategy. These include the Faster Cancer Treatment Indicators (58) and the Standards of

Service Provision for Breast Cancer Patients in New Zealand (59). These measures will

provide benchmarks for quality cancer care and are expected to improve care for all women

with breast cancer as well as reducing and hopefully eliminating inequities in access,

timeliness and quality of care along the symptomatic breast cancer care pathway.

The main strengths of this study include the completeness and comprehensive nature of the

study sample and data linkage with the national screening database which enabled us to define

and validate screen detected, interval and symptomatic non-interval cancers. Limitations of

this study include the absence of data on previous breast cancer screening behaviours of these

women, which is a factor known to be associated with cancer stage at diagnosis and long term

outcomes (137). Also, we were unable to ascertain the reasons for non-screening participation

in women with symptomatic non-interval cancer, which however is beyond the scope of the

present study.

In conclusion, we have observed a significantly higher rate of advanced stage breast cancers

among screening age Māori compared with NZ European women, of which approximately a

half was explained by lower rate of screen detected cancer among Māori women. Significant

differences in breast cancer survival by ethnicity and socioeconomic deprivation were

observed for non-screen detected, but not for screen detected breast cancer. Māori women who

do have screen detected breast cancers appear to do just as well as NZ European women

demonstrating the success of BSA for Māori women who are able to access this programme.

Achieving at least the national 70% biennial breast cancer screening rate for eligible Māori

women will likely make a significant contribution to reducing cancer deaths for Māori as well

as reducing inequities in cancer deaths between Māori and NZ European women in New

Zealand.

Results

131

5.4. Are there ethnic differences in breast cancer biology?

Preface:

This chapter contains an abbreviated version of a manuscript published in PLOS ONE.

Authors: Seneviratne S, Scott N, Shirley R, Kim B, Lawrenson R, Campbell I.

Title: Breast cancer biology and ethnic disparities in breast cancer mortality in New

Zealand: a cohort study

Journal: PLOS ONE


Impact factor: 3.73

Journal’s aims and scope: PLOS ONE is an open access peer-reviewed scientific

journal published by the Public Library of Science (PLOS) since 2006. It is the world’s

largest journal by number of papers published. It covers primary research from any

discipline within science and medicine. All submissions go through an internal and

external pre-publication peer review, but are not excluded on the basis of lack of

perceived importance or adherence to a scientific field. The PLOS ONE online

platform employs a "publish first, judge later" methodology, with post-publication user

discussion and rating features.

http://en.wikipedia.org/wiki/Open_access

http://en.wikipedia.org/wiki/Peer-reviewed

http://en.wikipedia.org/wiki/Scientific_journal

http://en.wikipedia.org/wiki/Scientific_journal

http://en.wikipedia.org/wiki/Public_Library_of_Science

http://en.wikipedia.org/wiki/Primary_research

http://en.wikipedia.org/wiki/Science

http://en.wikipedia.org/wiki/Medicine

http://en.wikipedia.org/wiki/Peer_review


132

Abstract:

Background:

Differences in cancer biological characteristics between Indigenous Māori and NZ European

women have been proposed as a possible contributor for higher breast cancer mortality in

Māori women. This study investigated differences in cancer biological characteristics and their

impact on breast cancer mortality disparity between Māori and NZ European women.

Methods:

Data on 2849 women with primary invasive breast cancers diagnosed between 1999 and 2012

were extracted from the WBCR. Differences in distribution of cancer biological characteristics

between Māori and NZ European women were explored adjusting for age and socioeconomic

deprivation in logistic regression models. Impacts of deprivation, stage at diagnosis and

biological characteristics on breast cancer mortality disparity between Māori and NZ European

women were explored in a Cox regression model.

Results:

Compared with NZ European women (n=2304), Māori women (n=429) had significantly

higher rates of advance staged and higher graded cancers. Māori women also had non-

significantly higher rates of ER/PR negative and HER-2 positive breast cancers. Higher odds

of advanced stage and higher grade remained significant for Māori after adjusting for age and

socioeconomic deprivation. Māori women had almost a 100% higher age and socioeconomic

deprivation adjusted breast cancer mortality risk compared with NZ European women (hazard

ratio=1.98, 1.55-2.54). Advanced stage at diagnosis and differences in breast cancer biological

characteristics explained a greater portion of the excess breast cancer mortality in Māori

compared with NZ European women (final hazard ratio=1.35, 1.04-1.75).

Conclusions:

More advanced cancer stage at diagnosis has the greatest impact while differences in

biological characteristics appear to be only a minor contributor for inequities in breast cancer

mortality between Māori and NZ European women. Underlying causes for these differences in

biological characteristics are unclear at present and is an area for future research.

Results

133

Background:

Advanced cancer stage at diagnosis in Māori has been shown to be the major contributor for

lower breast cancer survival in Māori compared with NZ European women (4). However,

significant ethnic differences in breast cancer survival remain after adjustment for stage at

diagnosis (4). Hence, factors other than stage including differences in timeliness and quality of

treatment and/or differences in cancer biology are likely to be important contributors to the

mortality disparity between Māori and NZ European women.

Data on biological differences in breast cancer between Māori and NZ European women have

so far been limited (70, 74, 169). The largest study to date was published by McKenzie et al

based on a cohort of women diagnosed during 1994-2004 from the New Zealand Cancer

Registry (70). The authors of this paper have reported significant differences in biological

characteristics, including higher rates of poorly differentiated and human epidermal growth

factor receptor type 2 (HER-2) positive cancers, and lower rates oestrogen (ER) and

progesterone receptor (PR) negative cancers in Māori compared with non-Māori/non-Pacific

(i.e. NZ European) women, which appeared to be independent of socioeconomic deprivation.

Two other groups from Auckland and Christchurch have also investigated biological

differences using smaller regional cohorts, but have reported on ethnic differences that

significantly differ from McKenzie at al report, including for tumour grade and hormone

receptor status (74, 169). Although all three studies have contributed significantly to the

knowledgebase on ethnic differences in biological characteristics, the exact nature of these

differences and their impact on breast cancer survival inequity between Māori and NZ

European women remain unclear at present.

We conducted this study to further investigate differences in breast cancer biological

characteristics between Māori and NZ European women, and to compare with previously

reported figures. We also attempted to identify the impact of biological differences on ethnic

disparities in breast cancer mortality in New Zealand.


134

Methods:

Study population:

All women with newly diagnosed invasive primary breast cancers from 01/01/1999 to

31/12/2012 were identified from the WBCR.

Study covariates:

Cancer stage at diagnosis was defined according to the Tumour, Node, and Metastasis (TNM)

staging system (267). Invasive tumour grade was defined according to the Elston and Ellis

modified Scarff-Bloom-Richardson breast cancer grading system (269). Oestrogen (ER) and

progesterone (PR) receptor status was determined based on results of immunohistochemistry

tests and classified as positive or negative. HER-2 status was based on Fluorescent In-Situ

Hybridization (FISH) test or when this was not available, on immunohistochemistry (270).


Categorical measures were summarized as numbers with percentages and continuous variables

were summarized as means with standard deviation. Chi squared (χ²) test for trend was used to

test for univariate differences in age adjusted rates of cancer biological characteristics between

Māori and NZ European women. Logistic regression models were used to explore associations

of tumour biology with socioeconomic deprivation and ethnicity, adjusting for age.

Multivariable Cox proportional hazard models were used to calculate hazard ratios with 95%

confidence intervals to identify the association of ethnicity, cancer stage and different cancer

biological factors with breast cancer specific mortality independently, and adjusting for age

and socioeconomic deprivation. Due to small numbers Pacific and Other ethnic group women

were excluded from analyses and ethnic comparisons were performed for Māori and NZ

European women. Breast cancer-specific survival curves for Māori and NZ European women

were estimated using the Kaplan-Meier method and compared by log-rank test. Deaths due to

causes other than breast cancer were considered as censored events. As some of the variables

included high numbers of missing data, survival analysis was repeated using women

diagnosed from 2006 onwards, where rates of missing data were significantly lower. A further

analysis was performed using only cases with complete data for all variables. Results of this

analysis were almost similar to those obtained from the full Cox proportional hazards

regression model, and these data are not presented in this report. Imputation of missing values

was not undertaken due to the similarity of these results.

Results

135

Results:

A total of 2856 women with new primary invasive breast cancer diagnosed in the Waikato

area over the study period were identified. Of these, Pacific (n=53) and Other (n=63) ethnic

women and seven women in whom a diagnosis of breast cancer was made post-mortem were

excluded, leaving 2733 for analysis. There were a total of 688 (25.2%) deaths, out of which

407 (59.2%) were due to breast cancer; 317 (77.9%) in NZ European and 90 (22.1%) in Māori

women. The study cohort was followed up for a median of 58 months (mean 66 months) and

67% women were followed up for a minimum of five years or until death.

Majority of the study women were of NZ European ethnicity (n=2304, 80.9%) and 15.1%

(n=429) were Māori. Distribution of tumour biological characteristics by ethnicity is shown in

Table 18. Māori women were significantly younger with a mean age difference of

approximately six years (61.5 vs. 55.6 years, p<0.001) keeping in with relatively younger

Māori population compared with NZ Europeans. A significantly higher age adjusted rate of

invasive ductal cancer was observed in Māori compared with NZ European women (85.0% vs.

80.5%, p=0.032). A corresponding reduction in the rate of invasive lobular carcinoma was

seen in Māori compared with NZ European (8.7% vs.11.7%, p=0.072), although this

difference was not statistically significant. Māori women had higher likelihoods of larger

breast tumours (p<0.001), positive lymphadenopathy (p<0.001), metastatic cancer (p<0.001)

and overall more advance staged cancer (p<0.001) compared with NZ European women.

Breast cancers among Māori were of higher grade (p=0.008) compared with NZ European

women, with less grade I and more grade II cancers after adjusting for age. Age adjusted rates

of ER+ and PR+ cancers tended to be lower (61.9% vs. 64.2%, p=0.373), and ER- and PR-

cancers tended to be higher in Māori (17.9% vs. 14.4%, p=0.071) compared with NZ

European women. When ER status is considered alone, age adjusted rate of ER positive

cancers was significantly lower in Māori compared with NZ European women (80.6% vs.

84.5%, p=0.011) (Appendix 4). Māori women had a statistically non-significant, higher age

adjusted rate of HER-2 amplified tumours (20.2% vs. 16.3%, p=0.068) compared to NZ

European women (Table 18).

As the rate of missing HER-2 data was relatively high (26.2%), an analysis was performed

only including breast cancers diagnosed from 2006, where the rate of missing HER-2 status

was only 4.3%. This showed figures similar to complete NZ European and Māori cohorts, with

a higher age adjusted rate of HER-2 positivity in Māori compared with NZ European women

(19.7% vs. 14.3%, p=0.076) (Table 18).


136

Table 18: Age and breast cancer biological characteristics at diagnosis compared between NZ

European and Māori women.

Characteristic NZ European (N=2304) Māori (N=429) p

n (crude %) Age

adjusted % n (crude %)

Age

adjusted %

Age

Mean ± SD 61.5 ± 13.9 55.6 ± 12.2 <0.001

T Stage

1 1249 (54.4) 53.2 178 (41.7) 42.5 <0.001

2 821 (35.8) 36.6 172 (40.4) 40.8

3 100 (4.4) 4.5 27 (6.3) 5.8

4 121 (5.3) 5.8 49 (11.5) 11.0

Unknown 13 3

N stage

0 1404 (61.4) 62.9 220 (51.9) 53.3 <0.001

1 582 (25.5) 24.7 130 (30.7) 29.7

2 188 (8.2) 7.8 40 (9.4) 9.3

3 112 (4.9) 4.6 34 (8.0) 7.7

Unknown 18 5

M Stage

0 2197 (95.4) 94.9 379 (88.3) 88.6 <0.001

1 107 (4.6) 5.1 50 (11.7) 11.4

Stage category

I 972 (42.2) 41.6 142 (33.1) 34.2 <0.001

II 887 (38.5) 39.1 163 (38.0) 37.5

III 338 (14.7) 14.2 74 (17.2) 16.9

IV 107 (4.6) 5.1 50 (11.7) 11.4

Histology

Ductal 1831 (81.2) 80.5 352 (85.2) 85.0 0.032a

Lobular 258 (11.4) 11.7 35 (8.5) 8.7 0.072b

Mixed 42 (1.9) 1.8 9 (2.2) 2.3

Other 125 (5.5) 5.9 17 (4.1) 4.0

Unknown 48 16

Results

137

Grade

Grade I 543 (25.3) 25.6 69 (17.6) 18.5 0.008

Grade II 1118 (52.1) 52.7 229 (58.4) 59.7

Grade III 485 (22.6) 21.7 93 (23.7) 21.8

Unknown 158 38

ER/PR

ER+/PR+ 1414 (64.1) 64.2 252 (60.9) 61.9 0.204

ER+/PR- 430 (19.5) 20.2 75 (18.1) 18.6 0.071c

ER-/PR+ 28 (1.3) 1.1 8 (1.9) 1.6

ER-/PR- 333 (15.1) 14.4 79 (19.1) 17.9

ER or PR

Unknown 99 15

HER-2

Negative 1239 (74.8) 75.2 257 (72.4) 73.8 0.091

Equivocal 135 (8.1) 8.5 21 (5.9) 6.0 0.069d

Positive 283 (17.1) 16.3 77 (21.7) 20.2

Unknown 647 74

TNBC e

No 1455 (91.5) 91.7 316 (92.9) 93.0 0.446

Yes 135 (8.5) 8.3 24 (7.1) 7.0

Unknown 714 89

Post 2005 NZ European (N=1247) Māori (N=278) p

n (crude %) Age


Age

adjusted %

HER-2

Negative 946 (79.5) 79.9 202 (74.8) 75.9 0.072

Equivocal 65 (5.5) 5.8 11 (4.1) 4.4 0.012d

Positive 179 (15.0) 14.3 57 (21.1) 19.7

Unknown 57 8

TNBC e

No 1054 (92.6) 92.6 238 (93.3) 93.5 0.536

Yes 84 (7.4) 7.4 17 (6.7) 6.5

Unknown 109 23

a ductal vs. other histology types, b lobular vs. other histology types, c ER and PR negative vs. other

receptor expressions, d HER-2 positive vs. negative/equivocal, e triple negative breast cancer.


138

Triple receptor status (ER, PR and HER-2) was determined for a total 1930 (70.7%) women

with invasive breast cancer. Of this group, 8.3% of cancers were negative for all three

receptors; i.e. triple negative breast cancer (TNBC). NZ European women had a higher age

adjusted rate of TNBC compared with Māori women, which was statistically not significant

(7.0% vs. 8.3%, p=0.446). Of women diagnosed from 2006 onwards, TNBC status was

available for 1393 (91.3%) women. Age-adjusted rate of TNBC was higher in NZ European

than in Māori, but this was still statistically non-significant (7.4% vs. 6.5%, p=0.532).

Increasing social deprivation significantly increased the risk of (age adjusted) advance staged

(p=0.001) and ER/PR negative (p=0.011) invasive cancers. No significant associations were

observed between deprivation and age adjusted rates of high tumour grade (p=0.095), HER-2

positivity (p=0.939) or TNBC status (p=0.270) (Table 19). Higher socioeconomic deprivation

status was significantly higher in Māori compared with NZ European women (Dep. 7-10

72.2% in Māori vs. 51.7% in NZ European, p<0.001, data not shown). Compared to NZ

European women, age and socioeconomic deprivation adjusted risk of advanced stage, higher

grade, ER and PR negativity and HER-2 positivity were higher while the rate of TNBC was

lower(8.3% vs. 7.0) in Māori women. Differences in stage and grade were statistically

significant, while differences in ER /PR, HER-2 and TNBC were not (Table 20).

Results from the survival analysis with Cox regression model are shown in Table 21. Māori

women had a significantly higher age adjusted breast cancer mortality compared with NZ

European women (HR 2.07, p<0.001) (data not shown). Adjusting for socioeconomic

deprivation marginally reduced the age-adjusted hazard of mortality for Māori compared with

NZ European from 2.07 (1.64-2.61) to 1.98 (1.55-2.54). As the proportion of screen detected

cancer was significantly higher in NZ European compared with Māori (37.4% vs. 30.1%

p=0.002, data not shown), detection method was included as a covariate in the survival model.

Adjusting for screening status and tumour stage (TNM stage) reduced the HR for mortality to

1.41 (1.09-1.82) (Appendix 5). Adjusting for tumour biological factors (i.e., grade, hormone

receptor status, HER-2 status and histology type) further attenuated this estimate (HR 1.42,

1.04-1.75). Further adjustments for treatment characteristics (i.e., chemotherapy and endocrine

therapy) and comorbidity resulted in a final hazard ratio of 1.25 (0.97-1.61), which was no

longer statistically significantly (p=0.088) (Table 21). Multivariate Cox regression model was

repeated only including women diagnosed from 2006 onwards, where rates of missing data

were significantly smaller (Table 21). Overall results of this model were much similar to the

model that included all women (final HR 1.25 vs. 1.28).

Results

139

Table 19: Age adjusted odds ratios (OR) for tumour biological characteristics by socio-economic deprivation category (NZDep 2006).

Deprivation

quintile

Stage a

n=2849

Grade b

n=2647

ER c

n=2784

PR d

n=2733

HER-2 e

n=2101

OR 95% CI p OR 95% CI p OR 95% CI p OR 95% CI p OR 95% CI p

Dep 1-2 Ref 0.001 Ref 0.183 Ref 0.007 Ref 0.579 Ref 0.961

Dep 3-4 0.84 (0.56-1.28) 0.78 (0.54-1.15) 1.07 (0.66-1.75) 0.91 (0.64-1.29) 0.88 (0.54-1.44)

Dep 5-6 0.84 (0.59-1.20) 1.05 (0.75-1.46) 1.43 (0.95-2.15) 0.97 (0.72-1.30) 0.94 (0.63-1.40)

Dep 7-8 1.22 (0.87-1.78) 0.99 (0.72-1.38) 1.47 (0.99-2.20) 0.87 (0.65-1.16) 0.90 (0.61-1.34)

Dep 9-10 1.35 (0.94-1.94) 1.17 (0.84-1.64) 1.90 (1.27-2.82) 1.04 (0.77-1.39) 0.99 (0.66-1.47)

a - Stage III & IV compared with stage I & II, b - Grade II & III compared with grade I, c - ER negative compared with positive, d - PR negative compared with positive, e -

HER-2 positive compared with HER-2 equivocal and negative

Table 20: Age and deprivation (NZDep 2006) adjusted odds ratios (OR) with 95% confidence intervals (95% CI) for breast cancer biological characteristics for Māori

compared with NZ European women.

Ethnicity

Stage a

n=2733

Grade b

n=2537

ER c

n=2669

PR d

n=2621

HER-2 e

n=2012

OR 95% CI p OR 95% CI p OR 95% CI p OR 95% CI p OR 95% CI p

NZ European Ref <0.001 Ref 0.004 Ref 0.482 Ref 0.400 Ref 0.154

Māori 1.63 (1.28-2.08) 1.52 (1.14-2.02) 1.13 (0.86-1.48) 1.13 (0.90-1.42) 1.24 (0.92-1.67)

a - Stage III & IV compared with stage I & II, b - Grade II & III compared with grade I, c - ER negative compared with positive, d - PR negative compared with positive, e -

HER-2 positive compared with HER-2 equivocal and negative.


140

Table 21: Cox regression model for factors associated with breast cancer specific mortality in

Waikato, New Zealand 1999-2012.

Characteristic Univariate Multivariate Multivariate

(Post 2005 only)

HR 95% CI p HR 95% CI p HR 95% CI p

Ethnicity a

NZ

European Ref <0.001 Ref 0.008 0.246

Māori 1.98 1.55-2.54 1.42 1.04-1.75 1.26 0.86-1.85

Year of diagnosis

1999-2002 Ref 0.556 Ref 0.034 -

2003-2006 1.13 1.19-1.43 1.12 0.86-1.47 Ref 0.872

2007-2009 0.95 0.70-1.27 0.78 0.55-1.11 0.97 0.63-1.47

2010-2012 0.97 0.66-1.41 0.84 0.56-1.27 0.88 0.52-1.46

Mode of detection

Non-screen Ref <0.001 Ref 0.007 Ref 0.045

Screen 0.26 0.20-0.35 0.67 0.48-0.90 0.52 0.28-0.98

T stage

T1 Ref <0.001 Ref <0.001 Ref <0.001

T2 3.33 2.56-4.33 1.91 1.45-2.51 3.33 1.80-6.18

T3 8.68 6.01-12.5 3.38 2.24-5.10 5.16 2.38-11.1

T4 18.6 13.8-25.1 3.35 2.33-4.82 4.95 2.41-10.1

N stage

N0 Ref <0.001 Ref <0.001 Ref <0.001

N1 2.03 1.58-2.60 1.61 1.23-2.09 1.36 0.86-2.16

N2+ 4.04 3.01-5.42 2.75 1.99-3.78 2.90 1.71-4.94

M Stage

M0 Ref <0.001 Ref <0.001 Ref <0.001

M1 15.4 12.2-19.4 4.16 3.08-5.62 5.61 3.64-8.63

Grade

I Ref <0.001 Ref <0.001 Ref <0.001

II 4.93 2.90-8.38 3.19 1.84-5.45 4.73 2.78-22.5

III 13.4 7.85-22.7 6.15 3.52-10.8 7.90 3.77-36.1

Results

141

ER/PR

ER/PR + Ref <0.001 Ref <0.001 Ref 0.009

ER & PR - 2.47 1.98-3.07 1.56 1.22-1.88 1.69 1.14-2.53

HER-2

Negative Ref <0.001 Ref 0.197 Ref 0.192

Equivocal 0.62 0.38-1.03 0.96 0.58-1.59 1.18 0.56-2.52

Positive 1.80 1.39-2.33 0.90 0.68-1.18 0.66 0.43-1.01

Histology

Ductal Ref 0.293 Ref 0.436 Ref 0.484

Lobular 0.87 0.62-1.22 0.91 0.63-1.29 0.31 0.18-0.52

Mixed 0.91 0.45-1.83 1.01 0.49-2.06 1.07 0.42-2.74

Other 0.58 0.32-1.05 0.69 0.37-1.29 0.49 0.22-1.07

HR - hazard ratios, 95% CI - 95% confidence intervals, a – adjusted for age and socio-economic

deprivation

Discussion:

From this study we have observed some differences in breast cancer biological characteristics

between Māori and NZ European women. Although Māori women had higher likelihoods of

exhibiting certain biological characteristics associated with worse breast cancer outcomes, this

appears to be only a minor contributor while advanced stage at diagnosis in Māori had the

greatest impact towards the breast cancer survival inequity between Māori and NZ European

women. Overall, Māori women had higher rates of advance staged and higher grade, and

possibly a higher rate of HER-2 positive cancers. No significant differences were observed in

rates of ER/PR negative or triple negative breast cancers (TNBC).

We have observed several key differences in our findings compared with previous studies (70,

74, 169). For example, McKenzie study based on the New Zealand Cancer Registry, reported

that Māori women have higher rates of ER/PR positive and poorly differentiated (i.e. grade

III) cancers compared with NZ European women (70). A higher rate of grade III cancers in

Māori was reported from a study based on the Auckland Breast Cancer Register (74), while a

third study from Christchurch reported Māori women to have a significantly lower rate of

grade III cancers compared with NZ European women (169). Differences in sample selection,

rates of missing data, statistical methods used for analysis and possible regional variations in

breast cancer could explain some of these differences. For instance, McKenzie study based on


142

the New Zealand Cancer Registry included very high rates of missing data; 65.1% for tumour

grade and 59.8% for ER status. Further, as the authors of the Christchurch study have

proposed (169), regional variations in breast cancer biological characteristics, may also have

contributed, especially for differences observed between Auckland and Christchurch datasets.

Such regional differences have been reported from the USA (86, 166), which could be related

to differences in distribution of risk factors associated with tumour biological expressions.

African American women with breast cancer in both the USA and the UK are known to

harbour more high grade, ER/PR negative and TNBC than their European American

counterparts (11, 165-167). Further, these differences are known to be major contributors for

excess breast cancer mortality in African American women (11, 166). Although, compared to

NZ European women, Māori women had a higher rates of ER/PR negative cancers (crude rate

15.1% vs. 19.1%) and grade III cancers (crude rate 22.6% vs. 23.7%), these rates were much

lower than rates of respective characteristics observed in African American women, in whom

the rates of ER/PR negative or grade III cancers were approximately 30-35% (86). Further, in

contrast to African American women, the rate of TNBC tended to be lower in Māori compared

with NZ European women, although this difference was not significant in either unadjusted or

adjusted analyses. It appears that differences tumour biology in Māori women may be a

contributor to higher mortality; though it certainly is not as significant a contributor as it is for

African American women.

Studies from the USA and the UK have demonstrated that women of lower socioeconomic

groups to have significantly higher rates of advance staged, ER/PR negative, high grade and

invasive ductal cancers compared with women living in affluent socioeconomic circumstances

(299, 300). Although many New Zealand studies have reported on the influence of

socioeconomic status on cancer stage at diagnosis for many cancers including breast (9, 122),

only the McKenzie study to date has reported on biological differences in breast cancer by

socioeconomic status (70). This study did not observe significant differences in tumour grade,

ER/PR and HER-2 status among different socioeconomic groups. In contrast, we observed a

higher age-adjusted rate of ER/PR negative cancers in women of low socioeconomic groups,

which was marked for women from the most deprived socioeconomic quintile. Further, in our

study, adjusting for age and socioeconomic status resulted only in a marginal attenuation of

higher grade cancers observed in Māori compared with NZ Europeans. Despite the differences

in the nature of biological differences between the two studies, both indicate that Māori

Results

143

women may have differences in breast cancer biology compared with NZ European women,

which are likely to be independent of age at diagnosis and socioeconomic deprivation.

Reasons for ethnic differences in breast cancer biology are largely unknown (301). It is

unlikely that there are innate genetic differences between ethnic groups resulting in more

aggressive breast cancers for some groups. Ethnic groups are not always from discrete

ancestry groups. For example, NZ European and Māori women descend from a wide range of

ancestry groups and each group shares common ancestry groups as Māori women will have

European ancestry and some NZ European women will have Māori ancestry. It is likely that

socio-environmental factors put minority, Indigenous and other socially disadvantaged women

at risk of having more aggressive breast cancers than privileged ethnic groups (302). Race

theory suggests that there is not enough genetic heterogeneity within the human species to

subdivide humans into groups with distinct genetic groups (303). However, socio-

environmental factors can influence epigenetics (304).

There were a few limitations in our analysis. First, some biological characteristics had

significant proportions of missing data. For example HER-2 data were missing for

approximately 25% of women, most of who were diagnosed prior to 2006 when HER-2 testing

was not routine in New Zealand. Further, even among women diagnosed post-2006, in

addition to missing HER-2 rate of 4.3%, a further 5% had an equivocal result, which may have

been a source of misclassification bias. Second, we acknowledge the possible differences in

analysis and reporting of biological characteristics by different laboratories (305). More than

95% of the pathology tests for cancers included in our study were performed by two

laboratories; one public and one private. These two laboratories have used similar equipment,

tests and reporting protocols over the time period and are expected to have had minimal

analysis and reporting variations.

In conclusion, we have observed Māori ethnicity and lower socioeconomic status to be

significantly associated with some breast cancer biological characteristics associated with

worse cancer outcomes. However, differences in tumour biological factors appear to be

contributing minimally, while delay in diagnosis in Māori appears to have a major impact on

the breast cancer mortality inequity between Māori and NZ European women. Strategies

aimed at reducing breast cancer mortality in Māori should focus on earlier diagnosis through

increasing screening coverage and other methods, which will likely have a greater impact on

minimizing the breast cancer mortality inequity between Māori and NZ European women.


144

Results

145

5.5. Are there ethnic differences in delay in surgical treatment?

Preface:

This chapter contains an abbreviated version of a manuscript published in Ethnicity & Health

Authors: Seneviratne S, Campbell I, Scott N, Coles C, Lawrenson R.

Title: Treatment delay for Māori women with breast cancer in New Zealand

Journal: Ethnicity & Health


DOI:10.1080/13557858.2014.895976

Impact factor: 1.20

Journal’s aims and scope: This is an international academic journal designed to meet

the world-wide interest in the health of ethnic groups. It embraces original papers from

the full range of disciplines concerned with investigating the relationship

between ’ethnicity’ and ’health’ (including medicine and nursing, public health,

epidemiology, social sciences, population sciences, and statistics). The journal also

covers issues of culture, religion, gender, class, migration, lifestyle and racism, in so

far as they relate to health and its anthropological and social aspects.


146

Abstract

Background:

Differences in delay in treatment have been shown to be a factor contributing to ethnic

inequities in breast cancer mortality. This study investigated differences in delay for surgical

treatment of breast cancer by ethnicity, and evaluated roles of health system, socio-

demographic and tumour factors towards these differences.

Methods:

A retrospective analysis of prospectively collected data included in the WBCR for cancers

diagnosed from 01/01/2005 to 31/12/2010 was done. Differences in proportions of women

with delays longer than 31 and 90 days were analysed by ethnicity, adjusting for covariates.

Results:

Approximately 95% (1449 out of 1514) of women with breast cancer diagnosed in the

Waikato over the study period were included. Of women undergoing primary surgery

(n=1264), 59.6% and 98.2% underwent surgery within 31 and 90 days of diagnosis

respectively. Compared with NZ European women (mean 30.4 days), significantly longer

delays for surgical treatment were observed among Māori (mean=37.1 days, p=0.005) and

Pacific women (mean=42.8 days, p=0.005). Māori women were more likely to experience

delays longer than 31 (p=0.048) and 90 days (p=0.286) compared with NZ European women.

Factors predicting delays longer than 31 and 90 days in the multivariable model included

public sector treatment (OR 5.93, 8.14), DCIS (OR 1.53, 3.17), mastectomy (OR 1.75, 6.60),

higher co-morbidity score (OR 2.02, 1.02) and earlier year of diagnosis (OR 1.21, 1.03).

Inequities in delay between Māori and NZ European women were greatest for women under

50 years and those older than 70 years.

Discussion:

This study shows that significant inequities in timely access to surgical treatment for breast

cancer exist in New Zealand, with Māori and Pacific women having to wait longer to access

treatment than NZ European women. Overall, a high proportion of women did not receive

surgical treatment for breast cancer within the guideline limit of 31 days. Urgent steps are

needed to reduce ethnic inequities in timely access to breast cancer treatment, and to shorten

treatment delays in the public sector for all women.

Results

147

Background:

Delays in diagnosis and treatment for many cancers, including breast cancer, are associated

with lower survival rates. A meta-analysis published by Richards and colleagues (198)

demonstrated a definite and strong relationship between delay in treatment and lower survival

from breast cancer. A delay of more than three months from onset of symptoms to initiation of

treatment was associated with a 12% lower five-year survival compared with a delay of less

than three months. A recently published study from the USA report that a delay of ≥60 days

compared with <60 days from diagnosis to initiation of treatment was associated with 66%

and 85% increased overall and breast cancer-related death rates respectively, among patients

with advanced breast cancer (206). Advances in breast cancer treatment have contributed

towards significant improvements in breast cancer survival over past few decades. Thus,

delays in receipt of treatment could be an important contributor to ethnic disparities in survival

among NZ women. Longer delays from diagnosis to treatment have been shown for Māori

compared to non-Māori for colon and lung cancers in NZ (181, 182).

BSA quality standards state that at least 90% of women should receive their first surgical

treatment within 20 working days of receiving their final diagnostic result. However, figures

from BSA in 2008 show that this target was achieved for only 57.7% of Māori women

compared with 71.2% of non-Māori women (132). Given the lack of standards and audit for

management of women with symptomatic non-screen detected cancers, the disparities in

treatment timeliness are likely to be even greater.

The aim of this paper is to identify differences in delay for surgical treatment of breast cancer

between ethnic groups in the Waikato and to evaluate the role of health system, socio-

demographic and tumour factors in ethnic inequities in breast cancer treatment.

Methods

All first primary incidental breast cancers, i.e. invasive and ductal in-situ cancers (DCIS),

diagnosed during the period of the study from 01 January 2005 through 31 December 2010,

were identified from the WBCR.

To assess time from diagnosis to surgery, we identified all women whose first treatment was

surgery, with a pre-operatively confirmed breast cancer diagnosis. Women were excluded

from analysis if they did not receive a pre-operative pathological diagnosis of breast cancer


148

(by a needle biopsy or punch biopsy), if their primary treatment was not surgery or if they

received neo-adjuvant therapy. Delay was calculated in days; from the day, a decision to treat

the breast cancer surgically was discussed with the patient, to the date of primary surgery. A

threshold of 31 days was used as the limit for the longest acceptable delay in access to surgical

treatment in keeping with the Faster Cancer Treatment Indicators (58) set by the New Zealand

Ministry of Health. Additional calculations were performed using a three month (90 days)

delay threshold, which has been shown to be associated with a significant survival

disadvantage (198). Several authors have used the 30 and 90-day benchmarks for treatment

delay (204, 306), so providing an opportunity for comparison.

Student’s T tests and Chi squared tests were used to test differences in delay among groups by

age, ethnicity, stage, mode of diagnosis and year of diagnosis. Mann–Whitney U test was used

to compare delay with non-parametric variables such as deprivation and distance from

residence. Multivariable logistic regression analyses were performed to determine variables

associated with delays of more than 31 and 90 days. All variables with p values of <0.50 in

univariate analyses were included in the multivariable model. A manual stepwise backward

selection procedure was used to select variables to be included in multivariable model. All p

values were two-sided and p values <0.05 were considered significant. Variables were retained

as predictors in multivariable models if p<0.05 or if they were considered to be of significant

clinical or population health importance. Odds ratios and 95% confidence intervals were

calculated to quantify the risk of a delay of more than 31 and 90 days associated with each

identified factor.

Results

During the period of the study a total of 1514 first primary incidental breast cancers (invasive

and DCIS) were reported from the Waikato. Sixty-five women (4.3%) did not consent to

participate in the WBCR and were excluded from this study. Data from 1449 women (95.7%)

were available for analysis.

Socio-demographic and tumour factors:

Table 22 shows the distribution of age and tumour stage by ethnicity. Māori and Pacific

women with breast cancer were significantly younger (p<0.001 and p=0.003 respectively) than

NZ European women in keeping with younger age structure of Māori and Pacific populations

Results

149

in NZ. Māori and Pacific women had more advanced staged cancers at diagnosis compared

with NZ European women. The proportion of DCIS and stage I cancers (cancers <20mm)

were significantly higher (p=0.041) among NZ European (n=569, 49.1%) compared with

Māori women (n=94, 41.6%).

The majority of women (58.3%) in this study belonged to higher deprivation groups (quintiles

4 and 5). Significantly higher proportions of Māori (n=173, 76.5%, p<0.001) and Pacific

women (n=24, 80.6%, p<0.001) were from deprivation quintiles 4 and 5 compared with NZ

European women (n=630, 54.4%). Overall, more than two-thirds (n=899, 71.2%) of women

were residing within 50km of their treatment facility and only 5.1% (n=64) were from highly

rural (>100 km) areas (Table 24). A significantly higher proportion of Māori were from highly

rural areas compared with NZ European women (9.3% vs. 4.1%, p<0.001).

Table 22: Distribution of age and tumour stage by ethnicity

Ethnicity Total NZ European Māori Pacific Other p

n=1449

(100%)

n=1159

(80.0%)

n=226

(15.6%)

n=30

(2.1%)

n=34

(2.3%)

n % n (%) n % n % n %

Age Mean 59.9 61.2 55.4 54.2 53.2 <0.001

<40 78 (5.4) 53 (4.6) 16 (7.1) 5 (16.7) 4 (11.8)

40-49 272 (18.8) 198 (17.1) 59 (26.1) 6 (20.0) 9 (26.5)

50-59 385 (26.6) 294 (25.4) 70 (31.0) 8 (26.7) 13 (38.2)

60-69 378 (26.1) 313 (27.0) 51 (22.6) 8 (26.7) 6 (17.6)

70+ 336 (23.2) 301 (26.0) 30 (13.3) 3 (10.0) 2 (5.9)

Stage I 491 (33.9) 404 (34.9) 69 (30.5) 5 (16.7) 13 (38.2) <0.001

II 457 (31.5) 359 (31.0) 79 (35.0) 10 (33.3) 9 (26.5)

III + IV 304 (21.0) 231 (19.9) 53 (23.5 14 (46.7) 6 (17.6)

DCIS 197 (13.6) 165 (14.2) 25 (11.1) 1 (3.3) 6 (17.6)


150

Breast cancer screening:

Compared to NZ European women (n=753, 65.0%), higher proportions of Māori (n=157,

69.5%) and Pacific women (n=20, 66.3%) belonged to the ‘screening age’ (45 – 69 years) as

defined by the National Breast Cancer Screening Programme (BreastScreen Aotearoa / BSA).

The majority of NZ European women (n=480, 63.7%) within the screening age group were

diagnosed through breast cancer screening. In comparison, less than half of the Māori women

(49.7%, n=78), within this age group were diagnosed through screening, a difference which

was statistically significant (p<0.001).

Delay in surgical treatment:

Overall, 59.6% (n=731) and 98.2% (n=1241) women underwent surgery within 31 and 90

days of diagnosis respectively (Table 23). Delays in surgical treatment for Māori (p=0.005)

and Pacific women (p=0.005) were significantly longer than for NZ European women. Māori

were 74% more likely to have a delay of over 3 months and 37% more likely to have a delay

of over 31 days compared with NZ European women. This difference was statistically

significant (p=0.048) for a delay longer than 31 days but was not significant for a delay longer

than 90 days (p=0.286) (Table 24).

Table 23: Delay from diagnosis to primary surgery (in days) by ethnicity

Overall

(n=1264)

NZ

European

(n=1023)

Māori

(n=186) p OR (95% CI)

Mean delay (days) 31.60 30.41 37.10 0.005 -

Delay >31 days 523 (41.4) 410 (40.2%) 89 (47.8%) 0.048 1.37 (1.00 – 1.88)

Delay >90 days 23 (1.8) 16 (1.6%) 5 (2.7%) 0.286 1.74 (0.63 – 4.80)

(OR – Odds Ratio, CI – Confidence Interval)

A significant positive correlation (p=0.003) was observed between a woman’s age at diagnosis

and a delay longer than 31 days with older women experiencing longer delays. This

correlation was significant for NZ European women (p=0.001) but was not for Māori women

(p=0.76). Among Māori women, higher proportions women younger than 50 years and older

than 70 years experienced a delay longer than 31 days compared to NZ European women

(Figure 22).

Results

151

Figure 22: Distribution of percentage of women with a delay longer than 31 days by ethnicity (Māori

vs. NZ European) and age category

Compared to women with invasive cancer, higher proportions of patients with DCIS

experienced a delay longer than 31 days (50.6% vs. 38.5%, p=0.009) and 90 days (3.5% vs.

1.6%, p=0.078). No significant differences in delay were observed among different stages of

invasive disease (p=0.78). Overall only 2.4% (n=31) women had a CCI of ≥1 and rates of CCI

≥1 were 2.2% (n=4) for Māori and 2.6% (n=26) for NZ European women. Increasing

comorbidity score showed a significant association with a delay longer than 31 days

(p=0.016).

A comparison of mean delay and proportion with a delay of longer than 31 days was

performed between women diagnosed through the BSA programme (n=516, 40.8%) and

outside of it (n=748, 59.2%). No significant differences in mean delay (mean 31.4 vs. 31.7

days, p=0.13) or proportion with a delay of >31 days were observed between these two groups

(41.9% vs. 41.0%, p=0.77). Women diagnosed through BSA had a higher likelihood of having

DCIS compared to women diagnosed outside of BSA (22.9% vs. 7.9%). Therefore, we

recalculated delay between BSA and non-BSA for women with invasive disease only. There

remained no significant difference in mean delay (p=0.12) or proportion with a delay of >31

days (p=0.84) between the two groups. A similar analysis was performed for women treated

within the public health system only (n=904). This demonstrated that a significantly higher

proportion of non-BSA compared with BSA diagnosed women had a delay longer than 31

days (55.7% vs. 47.3%, p=0.013). This difference in delay among public sector treated BSA

35% 35%39% 40%

47%

64%

49%

38%43%

67%

0%

10%

20%

30%

40%

50%

60%

70%

<40 40-49 50-59 60-69 70+

NZ European

Māori

Age (Years)


152

compared to non-BSA was greater and more significant for Māori women (37.3% vs. 59.6%,

p=0.004) while the difference was smaller and non-significant between non-BSA and BSA

treated NZ European women (49.3% vs. 54.5%, p=0.164).

Table 24: Univariate analysis of factors associated with a delay of >31 days and >90 days

Characteristic n (%) Delay >31 days Delay >90 days


Ethnicity

NZ European 1023 (80.9) 1.00 1.00

Māori 186 (14.7) 1.37 1.01 – 1.88 0.048* 1.74 0.63 - 4.80 0.286 a

Pacific 23 (1.8) 1.63 0.71 - 3.73 0.247* 2.86 0.36 - 22.5 0.318 a

Other 32 (2.5) 1.02 0.50 - 2.09 0.950* 2.03 0.26 - 15.8 0.499 a

Age category (yrs) 0.104 0.322

<40 68 (5.4) 1.15 0.66 - 1.99 1.41 0.27 - 7.44

40-49 238 (18.8) 1.00 1.00

50-59 345 (27.3) 1.10 0.78 - 1.54 0.69 0.20 - 2.39

60-69 355 (28.1) 1.08 0.77 - 1.52 0.40 0.09 - 1.68

70+ 258 (20.4) 1.57 1.10 - 2.25 1.49 0.48 - 4.62

BSA vs. non-BSA b 0.774 0.156

BSA 516 (40.8) 1.00 1.00

non-BSA 748 (59.2) 0.97 0.77 - 1.21 1.98 0.77 - 5.05

Stage 0.073 0.352

Stage 0 (DCIS) 172 (13.6) 1.51 1.07 - 2.14 2.46 0.81 - 7.42

Stage I 483 (38.2) 1.00 1.00

Stage II 438 (34.7) 0.96 0.74 - 1.26 1.27 0.45 - 3.52

Stage III + IV 171 (13.5) 0.98 0.68 - 1.39 0.80 0.17 - 3.91

Type of operation <0.001 <0.001

BCS 788 (62.3) 1.00 1.00

Mastectomy 476 (37.7) 1.67 1.33 - 2.10 6.15 2.27 - 16.7

Results

153

Facility type <0.001 0.032

Public 904 (71.5) 1.00 1.00

Private 360 (28.5) 0.17 0.13 - 0.24 0.11 0.01 - 0.83

Deprivation quintile 0.009 0.881

Dep 1-2 134 (10.6) 1.00 1.00

Dep 3-4 122 (9.7) 1.37 082 – 2.29 3.35 0.34 – 32.6

Dep 5-6 275 (21.8) 1.31 0.84 – 2.02 2.46 0.28 – 21.2

Dep 7-8 393 (31.1) 1.72 1.13 – 2.60 2.41 0.29 – 19.7

Dep 9-10 340 (26.9) 1.69 1.11 – 2.58 2.80 0.34 – 22.9

Distance from hospital 0.152 0.513

<10 km 356 (28.2) 1.00 1.00

10 – 50 km 543 (43.0) 0.91 0.69 – 1.19 0.58 0.22 – 1.51

50 – 100 km 301 (23.8) 0.97 0.71 – 1.32 0.52 0.16 – 1.70

>100 km 64 (5.1) 1.67 0.98 – 2.86 1.24 0.26 – 5.89

Charlson score 0.016 0.396

0 1233 (97.5) 1.00 1.00

1 14 (1.1) 2.61 0.87 – 7.84 4.23 0.53 – 33.8

≥2 17 (1.3) 3.48 1.22 – 9.95 -

Year of diagnosis <0.001 0.102

2005 216 (17.1) 1.00 1.00

2006 224 (17.7) 0.60 0.41 - 0.88 3.45 0.71 - 16.8

2007 214 (16.9) 0.91 0.62 - 1.33 0.50 0.05 - 5.58

2008 205 (16.2) 0.46 0.31 - 0.68 4.35 0.91 - 20.7

2009 222 (17.6) 0.41 0.28 - 0.61 1.47 0.24 - 8.86

2010 183 (14.5) 0.44 0.29 - 0.66 1.18 0.16 - 8.48

(OR – unadjusted Odds ratio, 95% CI – 95% Confidence interval, a - pairwise p value compared with

NZ European women, b BreastScreen Aotearoa)


154

Women undergoing breast conserving surgery (BCS) as the primary surgical intervention had

a significantly shorter mean delay compared with women undergoing mastectomy (28.7 vs.

36.4 days, p<0.001). The mastectomy group had significantly higher proportions of women

with delays longer than 31 and 90 days compared with the BCS group (Table 24). A

significantly higher proportion (p=0.003) of Māori (46.8%, n=87) underwent mastectomy as

primary surgical treatment compared with NZ European women (35.4%, n=362).

We also examined delay among women who received primary surgical treatment from the

private system (n=360) compared with the public system (n=904). Mean delay for surgery in

the private sector was much shorter compared with the public sector (20.1 vs. 36.2 days,

p<0.001). Similarly, a significantly lower proportion of women treated in the private sector

experienced delays of >31 and >90 days compared with public sector (Table 24). A

significantly higher proportion of NZ European women compared to Māori women (32.5% vs.

8.1%, p<0.001) received surgical treatment in the private sector. No significant differences in

mean delay were observed between NZ European and Māori within public (35.5 vs. 38.9 days,

p=0.18) or private sector (19.7 vs. 14.6 days, p=0.06). Within the public sector almost equal

proportions of NZ European and Māori women experienced delays longer than 31 days (52%

vs. 50.9%, p=0.96) and 90 days (2.3% vs. 2.9%, p=0.65).

A significant positive correlation was observed between increasing socio-economic

deprivation and delay (p=0.007) as well as a delay longer than 31 days (p=0.008). However,

this was not significant for a delay longer 90 days (p=0.620). A delay beyond 31 days was

more pronounced among women from deprivation quintiles 4 and 5 (Table 24). Increasing

deprivation showed a significant negative correlation (p<0.001) with proportion of women

accessing treatment from the private sector. Only 21.7% (n=70) women from the highest

deprived group (deprivation quintile 5) accessed private sector treatment compared with

44.1% (n=56) women from the least deprived group (deprivation quintile 1). When public

sector women are considered alone, no significant associations were observed between

deprivation and mean delay (p=0.223) or proportion of women with delays longer than 31

(p=0.226) or 90 days (p=0.972). Over the study period (2005 – 2010) a gradual, but significant

(p<0.001) overall decline was observed in the proportion of patients with delays of >31 days

(Table 24). However, despite these overall improvements, a higher proportion of Māori

compared with NZ European women continue to experience delays longer than 31 days

(Figure 23). For example in 2010, 42.4% Māori women experienced delays longer than 31

days compared to only 31.2% for NZ European women.

Results

155

Figure 23: Yearly trend in proportional delay longer than 31 days by ethnicity (Māori vs. NZ

European)

Multivariable logistic regression analysis identified treatment in the public sector, stage of

disease, earlier year of diagnosis, higher Charlson Comorbidity Score (only for 31 day

benchmark) and mastectomy as surgical treatment as factors significantly associated with

longer treatment delays at 31 and 90 day benchmarks, after controlling for age, socioeconomic

state, mode of diagnosis and distance from hospital (Table 25). A similar regression model

was performed using only women with a CCI score of zero which yielded results much similar

to Table 25.

Discussion

From this population-based study from the Waikato, we report that non-Indigenous New

Zealand women are more likely to receive surgical treatment for breast cancer before

Indigenous women. The main driver for this inequity in timeliness of access to breast cancer

treatment between Māori and non-Māori women is greater access to private treatment for non-

Māori women. It is known that women who access breast cancer treatment early have a greater

chance of surviving their cancer. This significant difference in timely access to treatment is

likely to contribute towards a higher breast cancer survival for non-Māori compared with

Māori women.

54.6%

40.7%

48.8%

32.9%28.8%

31.2%

50.0%42.9%

66.7%

41.4% 40.9%42.4%

0%

10%

20%

30%

40%

50%

60%

70%

2005 2006 2007 2008 2009 2010

NZ European

Maori

Year of diagnosis


156

Table 25: Multivariable logistic regression analysis of factors associated with a delay of >31 days and

>90 days adjusted for age, stage, deprivation score, mode of diagnosis and distance from hospital

(Only NZ European and Māori women (n=1219) included in the analysis)

Characteristic Delay >31 days Delay >90 days


Ethnicity 0.988 0.521

NZ European 1.00 1.00

Māori 1.00 0.70 – 1.43 1.42 0.48 – 4.17

Stage 0.002 0.041

Stage 0 (DCIS) 1.60 1.08 – 2.36 3.17 0.94 – 10.55

Stage I 1.00 1.00

Stage II 0.82 0.60 – 1.12 0.64 0.22 – 1.89

Stage III + IV 0.62 0.40 – 0.95 0.34 0.07 – 1.76

Type of operation <0.001 <0.001

BCS 1.00 1.00

Mastectomy 1.73 1.30 – 2.29 6.60 2.31 – 18.88

Facility type <0.001 0.029

Public 1.00 1.00

Private 0.16 0.11 - 0.20 0.10 0.01 – 0.79

Charlson score 0.009 0.976

0 1.00 1.00

≥1 2.02 1.18 – 3.44 1.02 0.22 – 4.78

Year of diagnosis <0.001 0.073

2005 1.00 1.00

2006 0.59 0.39 - 0.89 4.22 0.84 – 21.17

2007 0.95 0.62 - 1.45 0.54 0.05 – 6.06

2008 0.40 0.26 - 0.61 4.84 0.98 – 23.81

2009 0.32 0.21 - 0.49 1.38 0.22 – 8.58

2010 0.43 0.28 - 0.68 1.42 0.19 – 10.48

(OR – Odds ratio, 95% CI – 95% Confidence interval)

Results

157

We also found that less than two percent of women experienced a delay of more than three

months for primary surgical treatment of breast cancer. A delay beyond three-months from

diagnosis to treatment is known to result in lower survival for women with breast cancer

(198). However, approximately 40% of women experienced delays between 31 days to three

months and the clinical significance of this delay is unclear. McLaughlin and colleagues (206)

recently reported a higher overall and breast cancer related mortality for women with

advanced breast cancer (stage III and IV) who experienced a treatment delay of more than 60

days compared with a delay less than 60 days. As Māori women (as well as Pacific women)

are more likely to be diagnosed with advanced disease, even moderate delays in treatment are

likely to have a bigger impact on outcomes than they would for NZ European women (206).

Ability to access private sector care is dependent upon either financial affordability or

availability of health insurance, both of which are related to socioeconomic status of patients.

Socioeconomic status has also been shown to be associated with delay within the public sector

because of cost of travel, time off work and availability of social support. In this study, Māori

and Pacific women were highly over-represented in higher deprivation categories compared

with NZ European women, which is similar to the distribution of deprivation for Māori and

Pacific populations in New Zealand. Socioeconomic status is known to be a major

contributory factor toward inequities in breast cancer treatment (4, 307-309) and outcome in

many countries (13). High likelihood of ethnic minority and Indigenous women belonging to

lower socioeconomic status groups is a phenomenon observed worldwide and, inequities by

ethnicity persist after controlling for socioeconomic status. We observed a positive correlation

between deprivation and longer delay for treatment, which was most pronounced among

women from the two highest deprivation groups (deprivation quintiles 4 & 5). Low

proportions of women from high deprivation groups accessing treatment from the private

sector was the major contributor for this difference. Furthermore, almost 60% of women

included in this study were from the two highest deprivation quintiles compared with 46% for

the whole Waikato population (41). This discrepancy was larger compared to differences

between the total population deprivation profile and the profile for people with cancer reported

by the Ministry of Health (310).

In this study, we did not observe a significant difference in delay between Māori and European

women within the public sector. These findings are in contrast to several previous studies,

which have shown that non-Indigenous patients are advantaged within the New Zealand public

health system with superior quality of care and shorter delays for diagnosis and treatment (35,


158

181). Although a lack of inequities in timeliness to treatment observed in the public sector

could be attributed to a selection bias of a group of women with pre-operatively confirmed

diagnosis of breast cancer, which itself is an indicator of better quality care, it is unlikely to

fully explain this lack of difference. Delay in treatment represents only one facet of the

multifaceted care for women and disparities in all facets of care need to be assessed and

addressed to achieve equity in cancer treatment for all New Zealand women.

There is a stringent quality framework for the treatment of breast cancer detected through BSA

providers and performance is regularly audited (132). Despite the lack of an overall difference

in delay between BSA and non-BSA diagnosed women, a significant difference in delay was

observed between BSA and non-BSA women treated within the public sector. This highlights

the importance of audit and regular monitoring of performance within public sector hospitals

to minimize treatment delays. Faster Cancer Treatment Indicators (58) were introduced by the

Ministry of Health to supplement existing National Cancer Streams and Cancer Care

Guidelines. Cancer Care Coordinators were recently introduced into the public health system

to help cancer patients navigate through complex cancer care pathways that involve several

disciplines of care (311). These measures are expected to standardize care within the public

health system and thereby minimize delays and inequities associated with cancer treatment.

Additional procedures carried out to diagnose multi-centricity or multi-focality, which

preclude BCS or delays associated with planning and performance of immediate surgical

breast reconstruction could explain some of the delays observed among women undergoing

mastectomy (309). Non-tumour related factors such as age, ethnicity, area of residence and

socioeconomic status have been shown to be associated with a woman’s decision-making

process regarding mastectomy (312). Fear of complications of radiotherapy and difficulties in

attending radiotherapy that invariably follows BCS are some of the major concerns, which

may tilt the decision towards mastectomy instead of BCS among some women. Ethnic

minority women are more likely to reside in rural areas and are more likely to belong to low

socioeconomic groups and, these factors could influence the higher mastectomy rates observed

among such women. Higher rates of mastectomy in turn lead to longer delays further

confounding the understanding of association between ethnicity and delay.

Significant reductions in mean delay and the proportion of patients with a delay longer than 31

days have been achieved in the Waikato over the period from 2005 to 2010. It is likely that

increases in the surgical team providing breast cancer services and streamlining of the cancer

treatment pathway have helped to achieve these improvements. However, no significant

Results

159

reductions in the gap between Māori and NZ European women in relation to treatment delay

have been achieved. This is a major cause for concern in the context of increasing incidence

and mortality from breast cancer among Māori compared to non-Māori women in New

Zealand.

Limitations of our study include limited sample size and non-inclusion of details on clinical

decision-making process which may have influenced observed differences. We did not assess

the extent that body mass index (BMI) may have had on delay in this study. However,

previous studies on treatment delay for breast cancer have not shown a strong association

between BMI and delay (313). Surgical treatment of breast cancer is generally associated with

lower risks of major or life threatening peri-operative complications for obese women (BMI

>30) compared with surgical treatment of other cancers such as colon or lung (314, 315). This

provides a likely explanation for absence of a significant association and, we believe that

exclusion of BMI data is unlikely to influence the overall results of this study.

Inequities in delay in surgical treatment represents only one-step in the multi-step cancer care

pathway through which women with breast cancer have to navigate. There is every reason to

believe that Māori women are likely to encounter delays along each step of the cancer care

pathway from the first presentation to a primary care clinician or a screening programme to

completion of treatment. We highlight the need for improving the services in the public sector

where >70% of all women and >90% of Māori women received breast cancer care, in order to

reduce overall delays as well as to reduce ethnic inequities in delay.

Recent initiatives introduced by the Ministry of Health including Faster Cancer Treatment

Indicators and Cancer Care Coordinators are expected to enhance efficiency of delivery of

cancer care. Although these strategies may help to reduce delays within secondary and tertiary

health care institutions, they are unlikely to eliminate barriers faced by ethnic minority women

in access to primary care, transportation, health literacy and social support. Faster Cancer

Treatment Indicators are also unlikely to address other barriers to quality care, which have not

been clearly identified yet. While commending the initiatives implemented by the Ministry of

Health we highlight the importance of further research and initiatives targeting inequities in

cancer care between Māori and non-Māori women.


160

Results

161

5.6. Are there ethnic differences in delay in initiating chemotherapy and

radiation therapy?

Preface:

This chapter contains an abbreviated version of a manuscript published in BMC Cancer.

Authors: Seneviratne S, Campbell I, Scott N, Kuper-Hommel M, Round G,

Lawrenson R.

Title: Ethnic differences in timely adjuvant chemotherapy and radiation therapy for

breast cancer in New Zealand: A cohort study

Journal: BMC Cancer


DOI: 10.1186/1471-2407-14-839

Impact factor: 3.32

Journal’s aims and scope: BMC Cancer is an open access, peer-reviewed journal that

publishes articles on all aspects of cancer research, including the pathophysiology,

prevention, diagnosis and treatment of cancers. The journal also publishes articles on

molecular and cellular biology, genetics, epidemiology, and clinical trials.


162

Abstract:

Background:

Indigenous and/or minority ethnic women are known to experience longer delays for treatment

of breast cancer, which has been shown to contribute to ethnic inequities in breast cancer

mortality. This study examined factors associated with delay in adjuvant chemotherapy and

radiotherapy for breast cancer, and their impact on the mortality inequity between Indigenous

Māori and European women in New Zealand.

Methods:

All women with newly diagnosed invasive non-metastatic breast cancer during 1999-2012,

who underwent adjuvant chemotherapy (n=922) or radiation therapy (n=996) as first adjuvant

therapy after surgery were identified from the WBCR. Factors associated with delay in

adjuvant chemotherapy (60-day threshold) and radiation therapy (90-day threshold) were

analysed in univariate and multivariate models. Association between delay in adjuvant therapy

and breast cancer mortality were explored in Cox regression models.

Results:

Overall, 32.4% and 32.3% women experienced delays longer than thresholds for

chemotherapy and radiotherapy, respectively. Higher proportions of Māori compared with NZ

European women experienced delays longer than thresholds for adjuvant radiation therapy

(39.8% vs. 30.6%, p=0.045) and chemotherapy (37.3% vs. 30.5%, p=0.103). Rural compared

with urban residency, requiring a surgical re-excision and treatment in public compared with

private hospitals were associated with significantly longer delays (p<0.05) for adjuvant

therapy in the multivariate model. Breast cancer mortality was significantly higher for women

with a delay in initiating first adjuvant therapy (HR=1.45, 95% CI 1.05-2.01). Mortality risks

were higher, albeit non-significantly for women with delays in chemotherapy (HR=1.34, 95%

CI 0.89-2.01) or radiation therapy (HR=1.28, 95% CI 0.68-2.40).

Conclusions:

Indigenous Māori women appeared to experience longer delays for adjuvant breast cancer

treatment, which may be contributing towards higher breast cancer mortality in Māori

compared with NZ European women. Measures to reduce delay in adjuvant therapy may

reduce ethnic inequities and improve breast cancer outcomes for all women with breast cancer.

Results

163

Background:

Ethnic disparities in receipt of breast cancer care are well documented, and have been shown

to contribute towards worse breast cancer outcomes among Indigenous and/or minority ethnic

women (316, 317). Indigenous and/or minority ethnic women are more likely to experience

longer delays in initiation of treatment for breast cancer (313, 318, 319), which are known to

increase risks of breast cancer recurrence and mortality (198, 209, 210, 320).

A substantial reduction in breast cancer mortality has been observed in developed countries

over the last two decades, which has been attributed to earlier diagnosis with widespread use

of screening mammography and advances in breast cancer treatment (321). Timeliness of

instituting treatment is crucial in order to obtain the maximum potential benefit from these

new and advanced treatments. Two recent meta-analyses have shown a 6% and 15% increase

in relative mortality rate with each 4-week delay in initiating adjuvant chemotherapy (207,

208). Although timeline thresholds given in treatment guidelines are sometimes arbitrary and

controversial, longer delays for surgery, chemotherapy and radiation therapy have all been

proven to be associated with poorer breast cancer outcomes including higher risks of

recurrence and mortality (198, 207-210).

In New Zealand, longer delays experienced by Māori in the receipt of cancer care have been

reported for surgical treatment of lung cancer (182) and for receipt of adjuvant chemotherapy

for bowel cancer (181). To date, no data are available on delays in adjuvant therapy

experienced by New Zealand women with breast cancer or ethnic differences in the receipt of

such treatment.

We conducted this study to identify ethnic differences in delay in initiating adjuvant

chemotherapy or radiation therapy following surgical treatment for invasive breast cancer. We

also explored time trends in delays and impact of delay on breast cancer outcomes in this

cohort of women with breast cancer.

Methods

Study population:

All newly diagnosed invasive female breast cancers during the period from 01/01/1999 to

31/12/2012, were identified from the WBCR (n=2848) for this study. Of this, women with


164

metastatic cancer at diagnosis (stage IV) (n=166), women who did not undergo primary

surgery (n=114) and women who received neo-adjuvant therapy (n=87), were excluded.

Delay in adjuvant therapy:

To assess time gap from surgery to initiation of first adjuvant therapy (i.e. chemotherapy or

radiation therapy), all women with non-metastatic invasive breast cancer undergoing surgery

as primary breast cancer treatment modality were identified (n=2481). Chemotherapy was

considered as the first adjuvant therapy for all eligible women undergoing adjuvant

chemotherapy (n=922, 37.2%) and radiation therapy was considered as the first adjuvant

therapy for women undergoing radiation therapy without prior adjuvant chemotherapy (n=996,

40.1%). The time gap to adjuvant chemotherapy and radiation therapy was defined as number

of days from the most definitive operation for the breast cancer to the first administration of

chemotherapy or radiation therapy (319). The definitive surgical procedure at the primary site

captured the most invasive surgical procedure at the primary site and included excisional

biopsy, wide local excision and mastectomy. Women who had delays of more than 365 days

for either chemotherapy or radiation therapy were excluded.

A threshold of 60 days was used as the acceptable threshold delay for initiating chemotherapy,

based on evidence from three recently published papers. These include two meta-analyses

which have demonstrated 6% and 15% worse overall and disease free relative mortality rates

for each 4-week delay in initiating chemotherapy (207, 208). A third study from the USA,

which included more than 6000 women found significantly worse disease free survival for

women with stage II-III or triple negative or HER-2 positive cancers, who experienced delays

longer than 60 days (210). As some previous studies have used a 90-day threshold delay for

chemotherapy (320, 322-324), we performed additional analyses with a 90-day threshold for

chemotherapy. For radiation therapy, a 90-day threshold was used, which has conventionally

been used in the assessment of radiation therapy delay (209).

Data analysis:

Continuous variables were summarized as mean/median with standard deviation (SD).

Independent samples median test was used to test differences in continuous variables. Chi

squared tests (χ2) for trend was used to test differences in delay among groups including age,

ethnicity, stage, mode of diagnosis (screen detected or symptomatic) and year of diagnosis.

Multivariable logistic regression analyses were performed to estimate independent association

between above factors and delays in initiating adjuvant therapy. Separate Cox regression

Results

165

models were used to identify the association between breast cancer specific mortality and

delay (overall, chemotherapy and radiation therapy) adjusting for covariates.

Results:

This study included a total of 1918 women of whom 922 (711 NZ European and 153 Māori)

received chemotherapy and 996 (853 NZ European and 113 Māori) received radiation therapy

as first adjuvant therapy. The median time gap for initiating adjuvant chemotherapy was 49

days (mean 52.6, SD 21.3) and for adjuvant radiation therapy was 76 days (mean 81.4, SD

32.5). Māori women experienced significantly longer median delays compared with NZ

European women for both adjuvant chemotherapy (median delay 54 vs. 49 days, p=0.017) and

radiation therapy (median delay 83 vs. 75 days, p=0.046). Overall, 318 (31.9%) women

experienced a delay longer than 90 days to receive radiation therapy and the number of women

who did not receive chemotherapy within 60-day threshold was 301 (32.4%). A total of 619

(32.3%) women experienced a delay in receiving first adjuvant therapy. Five per cent (n=46)

women experienced a delay longer than 90 days for chemotherapy. A significantly higher

proportion of Māori women experienced a delay longer than 90 days compared with NZ

European women (8.7% vs. 4.2%, p=0.025)

Univariate analysis of factors associated with delay in receiving first adjuvant therapy,

chemotherapy and radiation therapy are shown in Table 26 and the multivariable logistic

regression in Table 27. Māori or Pacific ethnicity compared with NZ European ethnicity,

earlier year of diagnosis, requiring a re-excision following primary surgery, longer distance

from the tertiary care hospital and receiving surgical treatment from a public versus private

hospital were associated with significantly longer delays (p<0.05) for first adjuvant therapy in

both unadjusted and adjusted models. For chemotherapy, a significant inverse association

(p=0.048) was observed between stage and proportion with delays longer than 60 days with

the smallest proportion observed for stage III disease. Delays longer than threshold limits for

chemotherapy and radiation therapy were significantly associated with re-excisions after

primary surgery and treatment in public hospital in both univariate and multivariate models.

Distance from treatment facility was significantly associated with delay in radiation therapy

(p=0.021), but not for delay in chemotherapy (p=0.540). Delay for radiation therapy has

significantly reduced over time (p<0.001), while delays for chemotherapy have increased

during 1999-2009, although a decline is observed over 2010-2012.


166

Table 26: Univariate analysis of factors associated with delay in first adjuvant therapy a, delay in

radiation and delay in chemotherapy for women with newly diagnosed invasive breast cancer

Characteristic

Total

(N=1918)

Delay in first

adjuvant therapy a

Delay in radiation

therapy >90 days

Delay in chemotherapy

>60 days

n (%) n (%) p n (%) p n (%) p

Ethnicity

NZ European 1564 (81.5) 478 (30.6)

261 (30.6)

217 (30.5)

Māori 266 (13.9) 101 (38.0) 0.022 45 (39.8) 0.045 57 (37.3) 0.103

Pacific 38 (2.0) 19 (50.0) 0.013 6 (54.5) 0.101 13 (48.1) 0.057

Other 50 (2.6) 21 (42.0) 0.112 7 (35.0) 0.676 14 (46.7) 0.065

Age (years)

0.846

0.704

0.091

<40 117 (6.1) 38 (32.5)

9 (40.9)

29 (30.5)

40-49 433 (22.6) 130 (30.0)

43 (32.8)

87 (28.8)

50-59 583 (30.4) 193 (33.1)

89 (34.2)

104 (32.2)

60-69 494 (25.8) 165 (33.4)

94 (29.2)

71 (41.3)

70-79 212 (11.1) 70 (33.0)

60 (33.0)

10 (33.3)

80+ 79 (4.1) 23 (29.1)

23 (29.1)

0

Stage at diagnosis

<0.001

<0.001

0.048

I 800 (41.7) 228 (28.5)

176 (27.0)

52 (35.4)

II 772 (40.3) 291 (37.7)

112 (43.1)

179 (35.0)

III 346 (18.0) 100 (28.9)

30 (36.1)

70 (26.6)

Year of diagnosis

<0.001

<0.001

<0.001

1999-2002 393 (20.5) 166 (42.2)

105 (62.5)

61 (27.1)

2003-2006 579 (30.2) 181 (31.3)

107 (35.9)

74 (26.3)

2007-2009 452 (23.6) 146 (32.3)

50 (20.0)

96 (47.5)

2010-2012 494 (25.8) 126 (25.5)

56 (20.0)

70 (32.7)

Screening status

0.162

0.005

0.247

Non-screen 1110 (57.9) 372 (33.5)

171 (36.3)

201 (31.5)

Screen detected 808 (42.1) 247 (30.6)

147(28.0)

100 (35.3)

Deprivation

0.257

0.795

0.147

Dep 1-2 213 (11.1) 55 (25.8)

29 (28.7)

26 (23.2)

Dep 3-4 206 (10.7) 66 (32.0)

36 (34.3)

30 (29.7)

Dep 5-6 487 (25.4) 161 (33.1)

75 (29.8)

86 (36.6)

Dep 7-8 532 (27.7) 174 (32.7)

98 (32.7)

76 (32.8)

Results

167

Dep 9-10 480 (25.0) 163 (34.0)

80 (33.6)

83 (34.3)

Primary surgery

0.913

0.014

0.058

BCS 1318 (68.7) 425 (32.2)

260 (30.4)

165 (35.6)

Mastectomy 600 (31.3) 194 (32.3)

58 (40.8)

136 (29.7)

Re-excision

<0.001

<0.001

0.002

No 1659 (86.5) 491 (29.6)

250 (28.6)

241 (30.7)

Yes 259 (13.5) 128 (49.4)

68 (55.3)

60(44.1)

Distance from hospital

0.019

0.021

0.540

<10km 630 (32.8) 192 (30.5)

94 (28.1)

98 (33.1)

10-50km 740 (38.6) 221 (29.9)

107 (29.2)

114 (30.5)

50-100km 462 (24.1) 169 (36.6)

100 (38.9)

69 (33.7)

>100km 86 (4.5) 37 (43.0)

17 (43.6)

20 (42.6)

Surgical facility type

<0.001

0.009

0.001

Private 632 (33.0) 163 (25.8)

73 (25.8)

90 (25.8)

Public 1286 (67.0) 456 (35.5)

245 (34.4)

211 (36.8)

Charlson score

0.175

0.434

0.520

0 1677 (87.4) 534 (31.8)

265 (31.7)

269 (32.0)

1+ 241 (12.6) 85 (35.3)

53 (33.3)

32 (39.0)

a delay in radiation therapy longer than 90 days or delay in chemotherapy longer than 60 days

Adjusted multivariable logistic regression model identified year of diagnosis, re-excision and

surgical treatment facility type to be independently associated with delay in first adjuvant

therapy as well as for delay in chemotherapy and radiation therapy (Table 27). Overall, Māori,

Pacific and Other ethnicity were associated with higher likelihoods of delay for chemotherapy,

radiation therapy and for first adjuvant therapy, although this was statistically significant only

for delay in radiotherapy for Māori and delay in first adjuvant therapy for Pacific women in

the multivariable model. Sensitivity analysis with 90-day chemotherapy delay threshold

yielded similar results in the multivariable regression model (Appendix 6) with year of

diagnosis (OR=1.37, p<0.001), re-excision (OR=3.96, p=0.001) and public hospital care

(OR=4.89, p=0.001) showing significant associations. Māori women had a non-significantly

higher risk for a chemotherapy delay longer than 90 days (OR=1.41, p=0.291) compared with

NZ European women in this model.


168

Table 27: Multivariable model for factors associated with delay in first adjuvant therapy, delay in

radiation therapy longer than 90 days and delay in chemotherapy longer than 60 days a

Delay in first adjuvant

therapy b

Delay in radiation

therapy >90 days

Delay in chemotherapy

>60 days

OR 95% CI p OR 95% CI p OR 95% CI p

Ethnicity

NZ European Ref.

Ref

Ref

Māori 1.32 0.98-1.77 0.069 1.87 1.17-3.02 0.010 1.17 0.78-1.74 0.452

Pacific 2.05 1.03-4.08 0.041 2.47 0.54-11.2 0.242 1.87 0.82-4.26 0.136

Other 1.71 0.94-3.11 0.082 1.32 0.46-4.05 0.564 2.03 0.93-4.44 0.076

Year of diagnosis c 0.79 0.72-0.87 <0.001 0.49 0.42-0.56 <0.001 1.17 1.03-1.34 0.017

Re-excision 2.47 1.85-3.28 <0.001 3.55 2.32-5.44 <0.001 1.81 1.19-2.73 0.005

Surgical facility

type 1.53 1.22-1.93 <0.001 1.59 1.12-2.26 0.010 1.58 1.15-2.16 0.005

Distance from

hospital 1.10 1.02-1.18 0.024 1.23 1.09-1.39 0.001 1.03 0.92-1.15 0.654

a Adjusted for age, tumour stage, socioeconomic deprivation and comorbidity score, b Delay in

radiation therapy longer than 90 days or delay in chemotherapy longer than 60 days, c Year categories

as in Table 26.

Time trends in 60-day chemotherapy and 90-day radiation therapy delay by ethnicity is shown

in Figure 24. Higher proportions of Māori women have consistently experienced longer delays

for radiation therapy compared with NZ European women over the study period which were

significant during 2003-2006 and 2007-2009 periods (p=0.010 and p=0.012, respectively). The

reduction in radiation therapy delay has been greater for NZ European than for Māori over

1999-2009, which has resulted in a widening of disparity in delay between Māori and NZ

European, although this gap seems to have narrowed over last three year period of the study.

Higher proportions of Māori have experienced delays longer than 60 days for chemotherapy

over 1999-2009 period, but since has declined below the rate for NZ European women over

2010-2012. For delays in chemotherapy longer than 90 days (Figure 25), the highest

proportion was seen during 2007-2009 period (overall 10.9%, NZ European 9.7%, Māori

15.8%) and since has declined to 5.2% (NZ European 4.6%, Māori 6.8%) during 2010-2012.

Results

169

Figure 24: Time trends in delay in adjuvant radiation therapy longer than 90 days (Panel A) and

adjuvant chemotherapy longer than 60 days (Panel B) for invasive breast cancer in Waikato, New

Zealand 1999-2012

Figure 25: Time trends in delay in adjuvant chemotherapy longer than 90 days for invasive breast

cancer in Waikato, New Zealand 1999-2012

A survival analysis using a multivariable Cox regression analysis adjusting for covariates

showed a significantly higher breast cancer specific mortality risk (HR=1.45, p=0.024) among

women who experienced delays in first adjuvant therapy (Table 28). A sensitivity analysis

with a 90-day delay threshold for chemotherapy yielded a similar increased trend for breast

cancer mortality for women who experienced a delay for first adjuvant therapy (HR=1.29,

2.7%1.8%

9.7%

4.6%3.0%

8.6%

15.8%

6.8%

0%

5%

10%

15%

20%

1999-2002 2003-2006 2007-2009 2010-2012

NZEuropean

A


170

0.81-2.05, p=0.288), although this difference was no longer statistically significant. Delay in

chemotherapy (HR=1.34, p=0.157) and radiation therapy (HR=1.28, p=0.449) were also

tended to be associated with higher hazards of breast cancer mortality although these did not

reach statistical significance.

Table 28: Cox proportional models for breast cancer specific mortality by delay in first adjuvant

therapy, radiation therapy and chemotherapy

HR a b 95% CI p

Delay in first adjuvant therapy 1.45 1.05-2.01 0.024

Delay in radiotherapy >90 days 1.28 0.68-2.40 0.449

Delay in chemotherapy >60 days 1.34 0.89-2.01 0.157

a Performed in three separate Cox regression models, b For stage I-III invasive breast cancer adjusted

for age, ethnicity, stage of disease (i.e. tumour size and number of positive lymph nodes), tumour

grade, oestrogen receptor status, lympho-vascular invasion, year of diagnosis, comorbidity score and

receipt of adjuvant therapy)

Discussion:

From this study we found that among women with non-metastatic invasive breast cancer in the

Waikato, New Zealand, almost a third (32.3%) experienced a delay in initiating radiation

therapy or chemotherapy as first adjuvant therapy following primary surgical treatment.

Furthermore, Māori and Pacific compared with NZ European, rural compared with urban

dwelling women and women who received surgical treatment in public compared with private

hospitals had significantly higher likelihoods of experiencing delays longer than thresholds for

adjuvant therapy. Increasing socioeconomic deprivation tended to be non-significantly

associated with longer delays in adjuvant therapy while no association was observed between

delay and patient age. Although delay in radiation therapy seems to have improved over time,

substantial proportions of women continue to experience clinically significant delays for both

chemotherapy and radiation therapy.

Delays in adjuvant therapy for breast cancer experienced by disadvantaged populations

including minority and Indigenous ethnic groups are well documented (316, 317). From a

study based on over 100,000 women from the US National Cancer Database, Fedewa and

Results

171

colleagues reported that Black African women were 30% and 50% more likely to experience

delays longer than 60 and 90 days respectively, for initiation of adjuvant chemotherapy

compared with White European women (319). We observed a similar pattern where greater

proportions of Māori women experienced longer delays for chemotherapy (23% higher for 60-

day and 100% higher for 90-day delay) compared with NZ European women. However, from

2007-2009, there were no significant inequities between Māori and NZ European women and,

in 2010-2012 Māori women were less likely to experience a delay in accessing chemotherapy.

Timeliness in initiating treatment for breast cancer is of greater importance for women with

more advanced or more aggressive cancers (i.e. stage II or III, hormone receptor negative,

HER-2 positive) (206, 210). Māori women are more likely to be diagnosed with more

advanced disease and are more likely to have hormone receptor negative and HER-2 positive

breast cancers (4, 13), and hence longer delays in adjuvant therapy, as demonstrated in this

study are likely to have a greater impact on breast cancer survival in Māori women. However,

women with more advanced cancers seem to have had shorter delays, possibly due to

prioritized care for these higher risk women and, hence likely to have had a minimal

differential impact on higher mortality in Māori compared with NZ European women.

Delays in cancer adjuvant therapy are associated with factors including lack of access to

healthcare, difficulties with navigating the health system, geographic distance to treatment

facility, availability of transport and ability to take time-off work to attend adjuvant therapy

(316, 325, 326). Further, women of some ethnic minority populations including Black

Africans in the USA have been shown to be less willing to undergo adjuvant treatments

because of greater fear of side-effects and lack of knowledge on potential benefits (255, 327).

Longer delays for adjuvant therapy observed among Māori compared with NZ European

women in this study were likely due to a combination of these factors as Māori are more likely

to be socioeconomically deprived and live in rural areas with less access to transport compared

with NZ Europeans (34). These differences were observed despite temporary accommodation

and/or free transport been provided by the Waikato District Health Board for women requiring

these facilities to attend adjuvant chemotherapy and radiation therapy. Furthermore, the higher

risk of Māori compared with NZ European women for longer delays persisted even after

adjusting for deprivation and residence which probably was due to the impact of unmeasured

or under-measured confounders in the present study. For instance, Māori women are more

likely to have lower levels of education, lower health literacy and are less likely to have a

health insurance policy compared with NZ European women (34), factors which are known to


172

be associated with longer delays in cancer treatment (319). These factors were not included in

our analyses due to unavailability of these data from the WBCR.

We also observed that significantly smaller proportions of women who received surgical

treatment in the private sector had experienced delays beyond the threshold limits for both

chemotherapy and radiation therapy compared with women treated in the public sector. Ability

to afford treatment from private sector has a strong correlation with higher socioeconomic

status and/or having a health insurance policy. Hence, this observation supports, though

indirectly, affluent socioeconomic background and health insurance as factors influencing

shorter delays in the receipt of adjuvant therapy. This disparity is observed, despite more than

95% women included in the present study receiving their adjuvant chemotherapy and radiation

therapy free of charge from the public sector.

Rural residency is known to influence delay as well as a woman’s decision to undergo

adjuvant radiation therapy for breast cancer (328). Radiation therapy requires women to attend

a radiation facility five days a week over a period ranging from three to six consecutive weeks

and due to this many rural women prefer mastectomy over BCS (328). This is of greater

importance for the study women as the Waikato District Health Board covers an area of over

20,000 square kilometres and yet has provided radiation therapy services through a single

central facility. In comparison, chemotherapy in most instances requires only once in three

weeks and/or once a week visits to a chemotherapy facility and is less likely to be influenced

by rural residence. Consistent with this, no significant differences in delays for chemotherapy

were observed between urban and rural women in the present study.

Several effective strategies for minimizing delays in adjuvant therapy and reducing inequities

in delay between socioeconomic and ethnic groups are reported in literature. These include

improving access through increasing supply or efficient usage of existing cancer resources

through coordinated cancer care, decentralization of cancer services and through improving

patient health literacy (329-332). As we have observed, almost a third of women have

experienced delays in adjuvant therapy beyond the threshold limits and it appears that

overloading of oncology services was a likely factor. The greatest proportion of women

experiencing delays longer than three months for radiation therapy was seen during 1999-2002

which was a time when a severe nationwide shortage of radiation therapy services was

experienced in New Zealand (237). Since then the supply of radiation facilities has increased

resulting in a gradual reduction in proportion of women experiencing delays longer than three

months. However, the increase in supply of radiation therapy facilities has been inadequate or

Results

173

lagging behind to keep up with the increase in number of patients requiring radiation therapy.

As a result, even in 2010-2012, about 20% women were observed to have experienced delays

longer than three months for radiation therapy. Delays in chemotherapy appeared to have

worsened during 1999-2009 followed by an improvement in the last time period. These

different patterns of delay in chemotherapy and radiation therapy may reflect issues at national

level as well as local issues of service capacity compared with demand.

Increasing the supply of oncology services alone are unlikely to eliminate inequities in delay,

as disadvantaged women (i.e. ethnic minority, socioeconomically deprived, rural, etc.) will

still be more likely to be subjected to longer delays. Improved patient navigation through

cancer care coordinators (CCC) has been shown to help reduce delays, especially for women

who are at-risk for longer delays which include women of minority ethnicity and low

socioeconomic groups (224). The Waikato District Health Board has two full-time CCC’s

providing support for women with breast cancer since 2009. It is likely that CCCs have made a

major contribution towards the observed reduction in delay and reduction in inequities

between Māori and NZ European women observed during 2010-2012. The Ministry of Health

has also identified CCC’s as a key strategy to increase quality and reduce inequalities in

cancer care and since 2012 has provided funding for all District Health Boards to employ

CCC’s for management of common cancers in New Zealand (229).

This study did not examine the type, duration, dose or rates of completion of adjuvant

chemotherapy or radiation therapy which are also known to impact on the efficacy of these

adjuvant therapies (333, 334). Further, although we have observed several associations with

longer delays, we were unable to identify causes for these delays i.e. whether delays were due

to longer wait time for appointments or due to patients not attending appointments, additional

investigations required due to patient comorbidity, etc. Another limitation of this study was the

inclusion of only small numbers of Pacific and Other women. Although we have observed

longer delays among Pacific women small sample size prevented further analyses.

In conclusion, we have observed significantly longer delays experienced by Indigenous Māori

women, rural women and women receiving surgical care in the public sector. Although delays

in adjuvant therapy appear to have improved over the study period, it is concerning to note the

substantial proportion of women continuing to experience clinically significant delays for

adjuvant breast cancer therapy. Reducing delays through improvements in availability,

efficiency and access to oncology services will not only minimize ethnic inequities in delay

but may improve outcomes for all women with breast cancer in New Zealand.


174

Results

175

5.7. Are there differences in the use of adjuvant therapy for breast cancer

by ethnicity

Preface:

This chapter contains an abbreviated version of a manuscript submitted for publication in

Australian and New Zealand Journal of Public Health

Authors: Seneviratne S, Campbell I, Scott N, Lawrenson R.

Title: Ethnic differences in use of adjuvant therapy for breast cancer in New Zealand

Journal: Australian and New Zealand Journal of Public Health

Impact factor: 1.89

Journal’s aims and scope: The Australian and New Zealand Journal of Public Health

(ANZJPH) publishes peer-reviewed research into public health, relevant to researchers,

practitioners and policy makers. The Journal has a major focus on Australia and New

Zealand but articles from other countries are accepted provided that the implications

for Australia and New Zealand are addressed. Authors from Australia and New

Zealand are encouraged to locate their papers in the international literature. The

Journal is multidisciplinary and aims to publish methodologically sound research from

any of the academic disciplines that constitute public health.


176

Abstract:

Background:

Inequities in use of breast cancer adjuvant therapy by ethnicity and socioeconomic status are

well documented, and are known to be contributing to lower breast cancer survival among

women of minority ethnicity and lower socioeconomic status. We investigated ethnic and

socioeconomic inequities in use of adjuvant radiotherapy and chemotherapy in a cohort of

women with breast cancer in the Waikato, New Zealand.

Methods:

All women with newly diagnosed invasive breast cancer during 1999-2012 were identified

from the WBCR. Oestrogen (ER) and progesterone receptor (PR) negative tumours ≥10mm in

diameter and ER and/or PR positive tumours ≥20mm in diameter in women younger than 70

years were deemed eligible for the use of chemotherapy (N=1212). Use of radiotherapy was

considered for all women who underwent breast conserving surgery (BCS), and in women

who underwent mastectomy, if the tumour diameter was ≥50mm or ≥4 of the lymph nodes

were involved (N=1708).

Results:

Of the women deemed to be eligible based on criteria used, 836 (69%) and 1491 (87.3%)

women were observed to have received chemotherapy and radiotherapy, respectively. In the

multivariate model, significantly lower use of radiotherapy was associated with Māori

compared with NZ European ethnicity (OR=0.62, 95% CI, 0.39-0.98), comorbidity (OR=0.32,

95% CI, 0.23-0.43), distance from radiation facility (OR=0.87, 95% CI, 0.77-0.98),

mastectomy compared with BCS (OR=0.32, 95% CI, 0.19-0.56) and non-screen compared

with screen detection (OR=0.51, 95% CI, 0.34-0.77). No significant associations were

observed between chemotherapy use and ethnic or socio-demographic factors.

Conclusions:

Significantly lower use of adjuvant radiotherapy observed in Indigenous Māori and rural

women indicate the existing barriers to access radiotherapy for these women. Increasing

availability and improving access for radiotherapy, especially for women who are at high risk

due to ethnicity, geography or socioeconomic status need to be recognized as priorities.

Results

177

Background:

Differences in quality and timeliness in treatment of breast cancer, including differences in the

use of adjuvant therapy have also been reported to be important contributors for ethnic and

socioeconomic disparities in breast cancer survival (80, 322, 335).

Indigenous Māori in New Zealand are known to have lower access, receive inferior quality

cancer care and experience longer cancer treatment delays compared with NZ Europeans for a

variety of cancers. For instance, Māori patients have been reported to experience longer delays

for surgical treatment lung cancer (182), and a lower use of chemotherapy for bowel cancer

(181) compared with NZ European patients. To date, data are sparse on possible ethnic

differences in use, quality or timeliness of adjuvant therapy for breast cancer in New Zealand.

We hypothesized that Māori women were less likely to have received adjuvant chemotherapy

and/or radiation therapy compared with NZ European women, based on standard treatment

guidelines (44, 52), which might have contributed to higher breast cancer mortality in Māori

women. To answer this question, we analysed cancer treatment data from a regional,

population based sample of women with breast cancer diagnosed over a period of 14 years.

Rates of adjuvant chemotherapy and radiation therapy use by socio-demographic and tumour

characteristics were analysed individually, and adjusting for covariates, to identify

associations between use of adjuvant therapy, and ethnicity and socioeconomic status.

Methods:

Study population:

All women with newly diagnosed primary invasive breast cancers during the period from

01/01/1999 through 31/12/2012, were identified from the WBCR (n=2848). Of this, women

with metastatic cancer at diagnosis (n=166) and women who did not undergo primary surgery

(n=114) were excluded.

Use of adjuvant therapy:

Chemotherapy: Chemotherapy eligibility was considered only for women younger than 70

years. For oestrogen (ER) and progesterone (PR) receptor negative cancers, a maximum

tumour diameter of ≥10mm (n=276, 22.8%) was considered as the threshold for

chemotherapy. For ER and/or PR positive tumours, defining a threshold for chemotherapy was


178

complicated as this decision in most situations was based on multiple factors including lymph

node involvement, tumour grade, lympho-vascular invasion, HER-2 status, and more recently,

with Ki-67 and tumour genotyping (336). For ER and/or PR positive or unknown tumours

cancers, we considered ≥20mm maximum tumour diameter as the threshold for chemotherapy

(n=936, 77.2%) (44, 52). Of the women considered to be eligible for chemotherapy (N=1212),

women who received either adjuvant or neo-adjuvant chemotherapy were considered to have

received chemotherapy. We also performed a separate analysis with a different threshold for

ER and/or PR positive cancers. For this analysis, cancers were considered eligible only if one

or more of lymph node positivity, tumour grade >1 or lympho-vascular invasion were present

in addition to the tumour diameter of ≥20mm.

Radiation therapy: Women who were deemed to be eligible for radiation therapy (n=1708)

were identified based on following criteria. All women undergoing breast conserving surgery

without a completion mastectomy (n=1354, 79.3%) were considered eligible, and for women

undergoing a mastectomy, if the maximum tumour diameter was ≥50mm or if ≥4 lymph nodes

were positive for tumour metastasis (n=354, 20.7%) were considered eligible (44, 52).

Data analysis:

Categorical measures were summarized as numbers observed with percentages and Chi

squared tests (χ2) for trend were used to test differences in use of chemotherapy and radiation

therapy among groups categorized by age, ethnicity, stage, mode of diagnosis and year of

diagnosis. Multivariable logistic regression analyses were performed to identify factors

independently associated with use of adjuvant chemotherapy and radiation therapy.

Results:

Use of chemotherapy:

Of the women deemed eligible for chemotherapy, 836 (69%) women had received

chemotherapy. No significant differences in the rates of chemotherapy use were observed

between Māori and NZ European women (68.3% vs. 68.7%, p=0.916). Chemotherapy use was

significantly higher in younger women (p<0.001), in women with a zero comorbidity score

(p<0.001), for women surgically treated in private hospitals (p=0.002) and for women with

non-screen detected cancer (p=0.033) (Table 29). Increasing socioeconomic deprivation

tended to be associated with lower use of chemotherapy overall, and for Māori and NZ

Results

179

European women, although this was not statistically significant (p=0.402). As expected,

chemotherapy use was higher for cancers which were associated with adverse prognostic

features including size larger than 5cm, positive lymph node status, higher grade, lympho-

vascular invasion (LVI) and HER-2 positivity. Similar trends in the use of chemotherapy by

tumour characteristics were observed for Māori and NZ European women.


chemotherapy for invasive breast cancer a in the Waikato, New Zealand 1999-2012

Chemotherapy use

Characteristic Total population

(N=1212)

Total

(N=1212)

NZ European

(N=924)

Māori

(N=218) p

n % n % n % n %

Age (yrs.)

<0.001

<40 100 8.3% 90 90.0% 58 89.2% 18 90.0%

40-49 338 27.9% 276 81.7% 200 81.3% 56 81.2%

50-59 434 35.8% 309 71.2% 245 72.5% 53 68.8%

60-69 340 28.1% 161 47.4% 132 48.0% 22 42.3%

Deprivation

0.402

Dep 1-2 139 11.5% 100 71.9% 88 71.5% 8 80.0%

Dep 3-4 126 10.4% 89 70.6% 72 70.6% 12 75.0%

Dep 5-6 313 25.8% 216 69.0% 174 67.7% 29 70.7%

Dep 7-8 315 26.0% 213 67.6% 160 67.8% 39 61.9%

Dep 9-10 319 26.3% 218 68.3% 141 68.4% 61 69.3%

Surgical hospital type

0.002

Private 406 33.5% 303 74.6% 271 74.0% 21 80.8%

Public 806 66.5% 533 66.1% 364 65.2% 128 66.7%

Diagnostic type 0.033

Screen detected 407 33.6% 235 57.7% 195 57.9% 30 57.7%

Non-screen detected 805 66.4% 601 74.7% 440 75.0% 119 71.7%

Charlson score

<0.001

0 1056 87.1% 754 71.4% 582 70.5% 124 71.7%

1+ 156 12.9% 82 52.6% 53 53.5% 25 55.6%


180

Diagnosis year 0.003

1999-2002 269 22.2% 201 74.7% 160 73.7% 31 79.5%

2003-2006 374 30.9% 267 71.4% 216 72.5% 38 66.7%

2007-2009 292 24.1% 183 62.7% 128 60.7% 39 65.0%

2010-2012 277 22.9% 185 66.8% 131 66.2% 41 66.1%

Grade

<0.001

Grade I 168 13.9% 70 41.7% 62 43.4% 6 33.3%

Grade II 617 50.9% 409 66.3% 305 66.0% 83 66.9%

Grade III 395 32.6% 340 86.1% 254 85.8% 57 83.8%

Unknown 32 2.6% 17 53.1% 14 60.9% 3 37.5%

ER/PR status

<0.001

ER &/or PR + 926 76.4% 598 64.6% 457 64.3% 103 63.6%

ER & PR - 276 22.8% 233 84.4% 173 84.4% 46 83.6%

Unknown 10 0.8% 5 50.0% 5 62.5% 0

T stage

<0.001

T1 368 30.4% 234 63.6% 196 62.6% 29 69.0%

T2 692 57.1% 484 69.9% 364 71.4% 83 62.4%

T3 83 6.8% 64 77.1% 40 74.1% 19 82.6%

T4 62 5.1% 50 80.6% 31 77.5% 18 90.0%

N stage

<0.001

0 406 33.5% 223 54.9% 165 54.6% 37 50.0%

1 537 44.3% 381 70.9% 291 70.0% 72 74.2%

2+ 269 22.2% 232 86.2% 179 86.9% 40 85.1%

LVI

<0.001

Negative 783 64.6% 484 61.8% 362 61.0% 88 62.0%

Positive 429 35.4% 352 82.1% 273 82.5% 61 80.3%

HER-2

<0.001

Negative 658 54.3% 414 62.9% 314 63.2% 79 60.8%

Equivocal 48 4.0% 25 52.1% 17 48.6% 6 66.7%

Positive 219 18.1% 188 85.8% 134 86.5% 37 80.4%

Unknown 287 23.7% 209 72.8% 170 71.7% 27 81.8%

a ER and PR negative cancers ≥10mm and ER and/or PR positive cancers ≥20mm are included

Results

181

Multivariate analysis of factors associated with chemotherapy use is shown in Table 30. Age,

comorbidity score and adverse tumour characteristics remained significant while socio-

demographic factors including ethnicity and surgical hospital type were not significant.


chemotherapy for invasive breast cancer in the Waikato, New Zealand 1999-2012


Māori ethnicity 1.01 0.66-1.52 0.992

Age a 0.88 0.78-0.98 0.017

Year of diagnosis b 0.96 0.82-1.12 0.588

ER and/or PR positive 0.36 0.24-0.54 <0.001

Deprivation 0.97 0.86-1.08 0.558

Charlson score 0.31 0.21-0.46 <0.001

Hospital type 0.77 0.56-1.06 0.109

T stage 1.19 0.96-1.47 0.105

N stage 2.03 1.67-2.46 <0.001

Grade 2.01 1.61-2.49 <0.001

LVI 1.82 1.30-2.54 <0.001

HER-2 1.82 1.30-2.54 <0.001

a age categories as in Table 29, b year categories as in Table 29

An additional analysis was performed with a different chemotherapy threshold for ER and/or

PR positive cancers, considering these cancers as eligible for chemotherapy only if the cancer

had one or more of lymph node positivity, lympho-vascular invasion or tumour grade >1 in

addition to a maximum tumour diameter of ≥20mm. This analysis yielded results much similar

to the analysis in Table 29 (Appendix 7). According to new criteria, 1168 women were found

to be eligible, and of this 824 (70.5%) have received chemotherapy; 623 (70.1%) of NZ

European and 149 (70.3%) of Māori women. Multivariable logistic regression analysis showed

trends similar to Table 30 and, for Māori, the adjusted odds of chemotherapy use was 1.25

(0.85-1.87, p=0.258).


182

Use of radiation therapy:

Characteristics associated with use of radiation therapy are shown in Table 31. Overall,

radiation therapy was used for 1491 (87.3%) of the women deemed to be eligible for radiation

based on selection criteria. Radiation therapy use was non-significantly lower among Māori

compared with NZ European women (84% vs. 87.8%, p=0.138). Younger age at diagnosis,

lower deprivation, later year of diagnosis, surgical care in a private hospital, shorter distance

from the hospital, screen detection, undergoing BCS and adverse tumour characteristics

including higher grade, stage and positive lymph node status were significantly associated

with increased likelihoods of receiving radiation therapy. Although a similar increasing trend

in the use of radiation therapy were seen for Māori and NZ European women over each year

category, these respective rates were lower for Māori compared with NZ European women.

Table 31: Socio-demographic and tumour characteristics associated with use of adjuvant radiation

therapy for invasive breast cancer in the Waikato, New Zealand 1999-2012

Radiation therapy use

Characteristic

Total population

(N=1708)

Total

(N=1708)

NZ European

(N=1418)

Māori

(N=225)

n (%) n (%) n (%) n (%) p

Age (yrs.)

<0.001

<40 79 (4.6) 76 (96.2) 50 (96.2) 16 (94.1)

40-49 328 (19.2) 302 (92.1) 231 (93.9) 52 (86.7)

50-59 499 (29.2) 456 (91.4) 377 (92.0) 66 (90.4)

60-69 471 (27.6) 417 (88.5) 361 (90.5) 45 (75.0)

70-79 218 (12.8) 172 (78.9) 161 (78.9) 8 (72.7)

80+ 113 (6.6) 68 (60.2) 65 (60.7) 2 (50.0)

Diagnosis year

0.003

1999-2002 357 (20.9) 296 (82.9) 255 (83.9) 30 (76.9)

2003-2006 506 (29.6) 443 (87.5) 396 (88.0) 37 (86.0)

2007-2009 406 (23.8) 354 (87.2) 284 (87.4) 50 (84.7)

2010-2012 439 (25.7) 398 (90.7) 310 (91.4) 72 (85.7)

Deprivation

0.006

Dep 1-2 178 (10.4) 167 (93.8) 156 (94.0) 5 (83.3)

Dep 3-4 186 (10.9) 163 (87.6) 140 (87.5) 17 (94.4)

Results

183

Dep 5-6 414 (24.2) 364 (87.9) 318 (88.1) 37 (90.2)

Dep 7-8 491 (28.7) 423 (86.2) 356 (87.9) 54 (77.1)

Dep 9-10 439 (25.7) 374 (85.2) 275 (84.4) 76 (84.4)

Distance

0.005

<10km 546 (32.0) 489 (89.6) 409 (90.5) 54 (85.7)

10-50km 650 (38.6) 579 (89.1) 488 (88.1) 70 (85.6)

50-100km 428 (25.1) 364 (85.0) 310 (85.4) 49 (83.1)

>100km 74 (4.3) 59 (79.7) 38 (79.2) 16 (76.2)

Diagnostic type

<0.001

Screen detected 750 (43.9) 698 (93.1) 602 (93.3) 76 (90.5)

Non-screen 958 (56.1) 793 (82.8) 643 (83.2) 113 (80.1)

Hospital type

<0.001

Private 535 (31.3) 488 (91.2) 453 (91.5) 25 (86.2)

Public 1173 (68.7) 1003 (85.5) 792 (85.8) 164 (83.7)

Surgery type

<0.001

BCS 1354 (79.3) 1213 (89.6) 1031 (89.7) 143 (88.8)

Mastectomy 354 (20.7) 278 (78.5) 214 (79.9) 46 (71.9)

Grade

0.493

Grade I 441 (25.8) 381 (86.4) 341 (87.0) 27 (79.4)

Grade II 865 (50.6) 754 (87.2) 626 (88.0) 102 (82.9)

Grade III 371 (21.7) 331 (89.2) 257 (89.2) 56 (87.5)

Unknown 31 (1.8) 25 (80.6) 21 (77.8) 4 (100)

T stage

<0.001

T1 1020 (59.7) 915 (90.1) 790 (90.1) 98 (89.9)

T2 518 (30.3) 437 (84.4) 356 (85.8) 57 (76.0)

T3 100 (5.9) 79 (79.0) 54 (76.1) 20 (87.0)

T4 70 (4.1) 58 (82.9) 43 (84.3) 14 (77.8)

N stage

0.042

0 1029 (60.3) 900 (87.8) 762 (87.8) 106 (86.9)

1 363 (21.2) 321 (88.9) 266 (90.2) 45 (83.3)

2+ 316 (18.5) 266 (84.2) 215 (93.1) 36 (76.6)

Charlson score

<0.001

0 1456 (85.2) 1312 (90.1) 1111 (90.7) 151 (85.8)

1+ 252 (14.8) 179 (71.0) 134 (69.4) 38 (77.6)


184

Multivariable analysis of factors associated with radiation therapy is shown in Table 32. Māori

compared with NZ European ethnicity (OR=0.62, 95% CI, 0.39-0.98), older age (OR=0.78,

95% CI 0.70-0.87), longer distance from the radiation facility (OR=0.87, 95% CI, 0.77-0.98,

higher comorbidity score (OR=0.32, 95% CI, 0.23-0.43), mastectomy compared with BCS

(OR=0.32, 95% CI, 0.19-0.56) and non-screen compared with screen detection (OR=0.51,

95% CI, 0.34-0.77) were significantly associated with lower likelihoods of receiving radiation

therapy in this model.


radiotherapy for invasive breast cancer in the Waikato, New Zealand 1999-2012


Māori ethnicity 0.62 0.39-0.98 0.041

Age a 0.78 0.70-0.87 <0.001


Deprivation 0.94 0.83-1.08 0.394

Distance 0.87 0.77-0.98 0.024

Charlson score 0.32 0.23-0.43 <0.001

Surgery in public vs. private 0.76 0.53-1.09 0.144

Non-screen vs. screen detection 0.51 0.34-0.77 0.001

Mastectomy vs. BCS 0.32 0.19-0.36 <0.001

T stage 0.95 0.75-1.19 0.645

N stage 1.21 0.97-1.50 0.085


Further analyses were performed for women undergoing BCS and mastectomy separately.

These analyses confirmed that Māori were less likely to receive radiation following

mastectomy (OR=0.56, 0.25-1.25, p=0.157) and BCS (OR=0.72, 0.39-1.34, p=0.304),

although these differences were not statistically significant. Significantly lower likelihoods of

receiving radiation following both BCS and mastectomy were seen for women of older age

(OR=0.68, 95% CI 0.60-0.78 & OR=0.81, 95% CI 0.68-0.95 respectively), non-screen

compared with screen detected (OR=0.39, 95% CI, 0.27-0.54 and OR=0.70, 95% CI 0.29-

1.65, respectively) and for women with higher comorbidity scores (OR=0.27, 95% CI 0.19-

0.39 and OR=0.30, 95% CI 0.17-0.53, respectively).

Results

185

Figure 26 shows time trends in the use of chemotherapy and radiation therapy by ethnicity.

Chemotherapy use seems to have gradually declined for both Māori and NZ European women

from 1999-2002 to 2007-2009. Rates of radiation therapy use seem to be increasing gradually

in NZ European women while the rates have plateaued for Māori since 2003-2006.

Figure 26: Time trends in use of chemotherapy (Panel A) and radiotherapy (Panel B) by ethnicity

Discussion:

This study has shown that the use of adjuvant radiation therapy for breast cancer was

significantly lower in Māori compared with NZ European women based on accepted practice

guidelines (44, 52). No significant difference in the use of chemotherapy was observed

between Māori and NZ European women. Significantly lower use of radiation therapy was

also seen among rural compared with urban dwelling women and non-screen compared with

screen detected women. Overall, the use of radiation was lower than expected based on

guidelines (44, 52), and was substantially worse for post-mastectomy radiation (78.5%) than

for radiation following BCS (89.6%). Although use of radiation therapy seems to have

improved over time, even during 2010-2012 a substantial proportion of women (10%) were

observed to have not received radiation therapy.

Lower use of adjuvant chemotherapy for minority ethnic cancer patients are well documented

in the USA and include chemotherapy for breast, colon and lung among many other cancers

(316, 337). Not only that these patients have experienced lower use of adjuvant chemotherapy,


186

but on many occasions were subjected to longer delays and use of chemotherapy regimens not

in keeping with recommended guidelines (317, 338, 339). Similarly, lower use and longer

delays for adjuvant chemotherapy for bowel cancer in Māori compared with non-Māori

patients have supported the existence of similar ethnic disparities in New Zealand (181).

Despite that, we did not observe a significant difference between Māori and NZ European

women in the use of adjuvant chemotherapy for breast cancer, either in univariate or

multivariate models. Further, we have not analysed the use of recommended regimens of

chemotherapy or rates of completion of chemotherapy in the present study. Hence, although

we have not observed an ethnic disparity in overall adjuvant chemotherapy use, further

research is needed to investigate possible disparities in other areas of chemotherapy use

including rates of completion and use of recommended regimens.

Overall, use of radiation therapy fell short of recommended guidelines, and was significantly

lower for Māori compared with NZ European women (44, 52). Similar inequities in the use of

adjuvant radiation therapy for breast cancer have been reported from the USA, between

minority African American and White American women (83, 316). It appears that socio-

demographically disadvantaged women (i.e. Māori, rural residence and high socioeconomic

deprivation) had higher likelihoods of not receiving adjuvant radiation, while no such

differences were observed for chemotherapy. Differences in difficulty in access for radiation

therapy in comparison to chemotherapy might have at least partially been responsible for this

difference. Adjuvant radiation for the study population was provided through the central

radiation facility at the tertiary hospital in Hamilton, and radiation therapy required these

women to attend the radiation facility five days a week over a period, ranging from four to six

weeks. For women residing in remote and rural areas this would have posed an obvious

difficulty due to difficulties with and cost of travel. Many rural women with breast cancers

suitable for BCS opting for mastectomy due to these reasons is well documented in the

literature (340). Women of low socioeconomic groups also face similar difficulties due to

difficulties with transport, taking time off work or due to lack of support to care for

dependants, resulting in lower use of radiotherapy (341). Higher proportions of Māori live in

rural areas and are more likely to be socioeconomically deprived contributing to lower

radiation therapy use in Māori. However, lower use of radiation in Māori persisted even after

adjusting for these factors, which suggested an independent association between Māori and the

use of radiation therapy.

Results

187

Women with screen detected cancer were significantly more likely to have received radiation

therapy compared to women with non-screen detected women, a pattern seen following both

mastectomy and BCS. Diagnostic and treatment indicators for women diagnosed through BSA

programme are routinely measured and performance of each screening provider is regularly

audited against pre-established criteria. However, similar quality measures or audit processes

were non-existent for symptomatically detected cancer. This provides a likely explanation for

higher radiation therapy rates seen for screen detected cancer, despite these cancers carrying a

lower risk of local recurrence compared with non-screen detected cancer. This observation

highlights a failure of the healthcare system, where women with lower risk cancers have likely

been prioritized to receive treatment over women with higher risk cancers. Such inequities in

care are likely to further exacerbate inequities in breast cancer outcomes seen between Māori

and NZ European women, especially since Māori women have a significantly lower screening

coverage (39), and as a result, a lower proportion of screen detected cancer.

There were several limitations included in this study. First, although we observed differences

in adjuvant therapy among some groups of interest and several associations, we could not

ascertain exact causes for non-use (i.e. not referred, not seen by an oncologist or patient

declined) due to non-availability of these data. Selection of patients for chemotherapy is

complicated and is based on multiple factors including age, tumour size, grade, ER/PR, lymph

node status and lympho-vascular invasion. As a result, criteria used for selection of women s

eligible for chemotherapy were not absolute, especially for women with ER/PR positive

cancers. We did not observe major differences in distribution of these tumour characteristics

between Māori and NZ European women, and hence, these factors are unlikely to have

influenced selection for chemotherapy in a differential manner.

In conclusion, we observed significantly lower use of radiation therapy for Māori and rural

women, although similar disparities were not observed for chemotherapy. Difficulties in

accessing radiation therapy appeared to be a major contributor towards differences observed

by ethnicity, geographic location and socioeconomic status. Failures of the healthcare system

to providing equitable care were also evident by the discrepancy in radiation therapy seen

between screen and non-screen detected women. Increasing availability and improving access

for breast cancer adjuvant therapy for women who are at high risk of not receiving adjuvant

therapy due to ethnicity, geography or socioeconomic position need to be recognized as

priorities, which may help minimize breast cancer outcome inequities.


188

Results

189

5.8. Does patient adherence with treatment contribute to inequity?

Preface:

This chapter contains an abbreviated version of a manuscript published in The Breast

Authors: Seneviratne S, Campbell I, Scott N, Kuper-Hommel M, Kim B, Pillai A,

Lawrenson R.

Title: Adherence to adjuvant endocrine therapy: Is it a factor for differences in breast

cancer outcomes by ethnicity in New Zealand?

Journal: The Breast


DOI: 10.1016/j.breast.2014.11.011

Impact factor: 2.58

Journal’s aims and scope: The Breast is an international, multidisciplinary journal for

clinicians, which focuses on translational and clinical research for the advancement of

breast cancer prevention and therapy. The Editors welcome the submission of original

research articles, systematic reviews, viewpoint and debate articles, and

correspondence on all areas of pre-malignant and malignant breast disease, including:

Surgery, Medical oncology and translational medicine, Radiation oncology, Breast

endocrinology, Epidemiology and prevention, Gynaecology, Imaging, screening and

early diagnosis, Pathology, Psycho-oncology and quality of life, Advocacy, Supportive

and palliative care and Nursing.

http://dx.doi.org/10.1016/j.breast.2014.11.011

http://www.journals.elsevier.com/the-breast/editorial-board/

http://ees.elsevier.com/thebreast/


190

Abstract:

Background:

Despite the benefits of adjuvant endocrine therapy for hormone receptor positive breast

cancer, many women are non-adherent or discontinue endocrine treatment early. We studied

differences in adherence to adjuvant endocrine therapy by ethnicity in a cohort of New

Zealand women with breast cancer and its impact on breast cancer outcomes.

Methods:

We analysed data on all women (n=1149) with newly diagnosed hormone receptor positive,

non-metastatic, invasive breast cancer who were treated with adjuvant endocrine therapy in the

Waikato during 2005 to 2011. Linked data from the Waikato Breast Cancer Registry and

National Pharmaceutical Database were examined to identify differences by ethnicity in

adherence to prescribed adjuvant endocrine therapy and the effect of sub-optimal adherence on

cancer recurrence and mortality.

Results:

Overall, a high level of adherence of ≥80% was observed among 70.4% of women, which

declined from 76.8% to 59.3% from the first to fifth year of treatment. Māori women were

significantly more likely to be sub-optimally adherent (<80%) compared with European

women (crude rate 37% vs. 28%, p=0.005). In the adjusted model Māori women were still

significantly more likely to sub-optimally adhere to endocrine therapy than European women

(OR=1.51, 95% CI 1.04-2.17). Sub-optimal adherence was associated with a significantly

higher risk of breast cancer mortality (HR=1.77, 95% CI, 1.05-2.99) and recurrence

(HR=2.14, 95% CI, 1.46-3.14).

Conclusions:

Sub-optimal adherence to adjuvant endocrine therapy appeared to be a likely contributor for

breast cancer mortality inequity between Māori and European women. Urgent research is

needed to identify effective ways to increase adherence to endocrine therapy, especially for

Māori women.

Results

191

Background:

Adjuvant endocrine therapy forms an integral part in breast cancer treatment and has shown to

reduce mortality from hormone receptor positive breast cancer by about 30% (49, 342).

Traditionally, endocrine therapy [tamoxifen or an aromatase inhibitor (AI) as single agent or

in sequence] was prescribed for 5 years, although recent studies have shown additional

improvement of breast cancer specific survival by continuing tamoxifen beyond 5 years (51).

Despite proven benefits, many women either do not take their medication daily as prescribed

(i.e. low adherence) or do not complete the full duration of treatment (i.e. discontinuation) for

the minimum of 5 years (343, 344). Based on previous studies, up to 22% of women

discontinue endocrine therapy before the end of first year of therapy and only about 50%

complete the full 5-years, while maintaining an optimum level of adherence (217, 345, 346).

These studies have also shown higher risks of breast cancer recurrence and mortality in

women who are sub-optimally adherent or who discontinue their treatment (344, 347).

This study was conducted to estimate the degree of adherence to adjuvant endocrine therapy

and to investigate ethnic, socio-demographic, tumour and treatment related factors associated

with poor adherence among women with hormone receptor positive breast cancer in New

Zealand. We also investigated the association between sub-optimal adherence and breast

cancer outcomes to determine the impact of adherence on ethnic inequities in breast cancer

outcomes.

Methods:

Study population:

All women newly diagnosed with invasive breast cancer from 01/01/2005 to 31/12/2011 were

identified from the WBCR (n=1558). Of this, 1207 women with hormone receptor positive,

non-metastatic (stage I to III), first primary breast cancer, who received adjuvant endocrine

therapy were identified from the national pharmaceutical database. All women with at least

one prescription for endocrine therapy after the date of primary surgery were deemed to have

started endocrine therapy. Of this group, a further 57 women who were started on endocrine

therapy not as adjuvant treatment, but as treatment following development of local or

metastatic recurrence, or first endocrine therapy prescription issued later than a year after the


192

date of diagnosis were excluded. The remaining 1149 women were analysed for treatment

adherence and outcomes.

Study covariates:

Follow-up duration was calculated from the first dispensing date of adjuvant endocrine

treatment to date of death or to date of last follow up when the patient was known to be alive

(censored on 31/12/2013).

Treatment adherence:

Prescription records for tamoxifen and AIs (i.e. anastrozole, letrozole and exemestane) for

each eligible woman for the period from 01/01/2005 to 31/12/2013 were obtained from the

National Pharmaceutical database. Prescription records were linked through the National

Health Index number, which is a unique identifier that is used to identify individuals within

the New Zealand health system. Dispensing date, drug type and number of days covered by

each prescription were recorded. An adherence index / medication possession ratio (MPR) for

each woman was calculated by dividing the number of days covered by prescriptions, by the

total number of days for the follow up period, until death, or up to 5 years. Any gaps in

treatment for more than 180 days were considered as discontinuation of therapy (347) and

were censored at the last date covered by final prescription prior to discontinuation. An

adherence index (MPR) of ≥80% was considered as a high/optimal level of adherence, which

is a figure widely used in previous literature (217, 347). Adherence indices were calculated

separately for each year of follow up and for the total follow up, to a maximum of 5 years.


Chi squared (χ²) tests for trend was used to test for univariate differences in distribution of

treatment adherence and outcomes among groups of interest. Factors associated with

adherence were explored in a multivariable logistic regression model. Multivariable Cox

proportional hazard models were used to calculate hazard ratios with 95% confidence intervals

to identify the association of adherence with breast cancer specific mortality, all-cause

mortality and cancer recurrence. Kaplan-Meier survival curves were used to calculate 5-year

crude breast cancer specific and disease free survival rates associated with high and sub-

optimal adherence.

Results

193

Results:

Median age of the cohort (n=1149) was 60 years (range 24-99). Median ages of NZ European

and Māori were 62 (range 24-99) and 57 (range 28-89) years, respectively. Median follow-up

duration was 51 months (inter-quartile range 32.1-73.0) months. There were a total of 131

(11.4%) cancer recurrences and 164 (14.3%) deaths, out of which 77 (47%) were due to breast

cancer. Overall, 51% of women were followed up for at least 5 years or until death.

A total of 509 (42.2%) women were started on tamoxifen and 698 (57.8%) were started on an

AI. Tamoxifen was the only endocrine therapy received by 269 (23.4%), while 521(45.3%)

women received AIs alone. Sequential therapy with tamoxifen and AIs was received by 359

(31.2%) women.

Overall, a high level of adherence (MPR ≥80%) was observed in 809 (70.4%) of women over

the total duration of therapy. Highest adherence was seen during the first year of therapy

where 76.8% maintained a high level of adherence. This figure gradually declined to 73.5%,

71.4%, 66.3% and 59.3% over the second to fifth years of treatment, respectively. Māori

women were observed to have a significantly lower adherence compared with NZ European

women, overall (62.1% vs. 72.5%, p=0.004) and over each year of treatment (Figure 27).Rates

of sub-optimal adherence were significantly higher (p<0.05) for Māori and Pacific women

compared with NZ European women, in both univariate and multivariable models.

Figure 27: Annual rates of high level of adherence (MPR≥80%) with adjuvant endocrine therapy for

hormone receptor positive invasive breast cancer for NZ European and Māori women

79.6% 76.1%73.7%

68.1%

60.2%64.8% 62.6% 60.8%

56.3% 54.9%

0%

20%

40%

60%

80%

100%

1st year 2nd year 3rd year 4th year 5th year

NZ European

Māori


194

Table 33: Factors associated with adherence to adjuvant endocrine therapy for hormone receptor

positive invasive breast cancer unadjusted and adjusted multivariable models

Adherence

≥80%

Adherence

<80% Unadjusted Adjusted

n (%) n (%) OR 95% C.I. p OR 95%

C.I. p

Ethnicity

NZ European 665 (82.2) 252 (74.1) 1.00 1.00

Māori 113 (14.0) 69 (20.3) 1.61 1.16-2.25 0.005 1.51 1.04-2.17 0.028

Other 21 (2.6) 5 (1.5) 0.63 0.23-1.68 0.356 0.55 0.20-1.54 0.256

Pacific 10 (1.2) 14 (4.1) 3.69 1.62-8.42 0.002 3.16 1.29-7.75 0.012

Age group (years)

<40 23 (2.8) 30 (8.8) 2.17 1.18-4.01 2.22 1.16-4.27

40-49 130 (16.1) 78 (22.9) 1.00 <0.001 1.00 0.015

50-59 203 (25.1 95 (27.9) 0.78 0.54-1.13 0.85 0.51-1.40

60-69 232 (28.7) 79 (23.2) 0.57 0.39-0.83 0.62 0.33-1.15

70-79 121 (15.0) 34 (10.0) 0.47 0.29-0.75 0.52 0.26-1.06

80+ 100 (12.4) 24 (7.1) 0.40 0.24-0.68 0.37 0.17-0.83

Deprivation decile

Dep 1-2 99 (12.2) 34 (10.0) 1.00 0.422 1.00 0.814

Dep 3-4 81 (10.0) 31 (9.1) 1.11 0.63-1.97 0.88 0.48-1.62

Dep 5-6 191 (23.6) 84 (24.7) 1.28 0.80-2.04 1.13 0.69-1.85

Dep 7-8 249 (30.8) 96 (28.2) 1.12 0.71-1.77 1.02 0.62-1.67

Dep 9-10 189 (23.4) 95 (27.9) 1.46 0.92-2.32 1.18 0.71-1.94

Residence profile

Urban 425 (52.5) 167 (49.1) 1.00 0.559 1.00 0.655

Semi-urban 318 (39.3) 142 (41.8) 1.14 0.87-1.48 1.15 0.86-1.54

Rural 66 (8.2) 31 (9.1) 1.20 0.75-1.90 1.05 0.63-1.73

Charlson score

0 682 (84.3) 294 (86.5) 1.00 0.107 1.00 0.171

1-2 116 (14.3) 37 (10.9) 0.74 0.50-1.10 0.90 0.58-1.39

3+ 11 (1.4) 9 (2.6) 1.90 0.78-4.63 2.49 0.90-6.87

Tumour stage

T1 444 (54.9) 187 (55.0) 1.00 0.342 1.00 0.198

T2 303 (37.5) 116 (34.1) 0.91 0.69-1.20 1.02 0.73-1.42

Results

195

T3 31 (3.8) 18 (5.3) 1.38 0.75-2.53 1.76 0.90-3.43

T4 28 (3.5) 19 (5.6) 1.61 0.88-2.96 2.02 0.98-4.15

Unknown 3 (0.4) 0

Lymph node stage

N0 478 (59.1) 203 (59.7) 1.00 0.597 1.00 0.685

N1 227 (28.1) 85 (25.0) 0.88 0.65-1.19 0.85 0.60-1.19

N2 100 (12.4) 50 (14.7) 1.18 0.81-1.72 1.00 0.62-1.61

Unknown 4 (0.5) 2 (0.6)

Grade

I 205 (25.3) 110 (32.4) 1.00 0.082 1.00 0.038

II 441 (54.5) 175 (51.5) 0.74 0.55-0.99 0.71 0.52-0.98

III 134 (16.6) 46 (13.5) 0.64 0.43-0.96 0.55 0.34-0.89

Unknown 29 (3.6) 9 (2.6)

Therapeutic Surgery

No 36 (4.4) 14 (4.1) 1.00 0.801 1.00 0.516

Yes 773 (95.6) 326 (95.9) 1.08 0.58-2.04 0.73 0.28-1.91

Chemotherapy

No 569 (70.3) 228 (67.1) 1.00 0.272 1.00 0.317

Yes 240 (29.7) 112 (32.9) 1.16 0.89-1.53 0.81 0.53-1.23

Radiotherapy

No 246 (30.4) 95 (27.9) 1.00 0.404 1.00 0.873

Yes 563 (69.6) 245 (72.1) 1.13 0.85-1.49 0.97 0.70-1.35

A significant trend (p<0.001) was observed between age and sub-optimal adherence with the

lowest rate observed in women over the age of 80 years (OR=0.37, 95% CI 0.17-0.83) and the

highest rate in women younger than 40 years (OR=2.22, 95% CI 1.16-4.27) (Table 33). There

was a trend for higher rates of suboptimal adherence among women of higher deprivation

categories, especially in the unadjusted model, which however was not statistically significant.

Thirty-four (10%) and 43 (5.3%) breast cancer deaths were observed among women with sub-

optimal (n=340) and high adherence (n=809), respectively. In both unadjusted and adjusted

Cox regression models, sub-optimal adherence was associated with a significantly higher risk

of breast cancer mortality (HR=1.62 and 1.77, respectively) and breast cancer recurrence

(HR=1.90 and 2.14, respectively) (Table 34 & Table 35). Adjusting only for adherence


196

reduced the hazard ratio for breast cancer mortality for Māori from 1.44 (95% CI 0.82-2.55) to

1.36 (95% CI 0.77-2.42). Hazard ratios for overall mortality were also higher among sub-

optimally adherent women (unadjusted HR=1.02, 95% CI 0.74-1.41, p=0.968, adjusted

HR=1.17, 95% CI 0.81-1.69, p=0.401), although these were statistically non-significant.

Table 34: Adherence to adjuvant endocrine therapy and breast cancer mortality unadjusted and

adjusted for age, comorbidity, deprivation, tumour factors (size, lymph node status, grade) and other

treatment modalities (surgery, radiotherapy and chemotherapy)

Unadjusted Adjusted

HR 95% C.I. p HR 95% C.I. p

Adherence

≥80% 1.00 0.036 1.00 0.033

<80% 1.62 1.03-2.54 1.77 1.05-2.99

Ethnicity


Māori 1.44 0.82-2.55 0.207 1.25 0.65-2.38 0.506

Other 0.66 0.09-4.79 0.684 0.60 0.08-4.53 0.620

Pacific 1.29 0.32-5.29 0.721 0.48 0.11-2.20 0.347

Table 35: Adherence to adjuvant endocrine therapy and breast cancer recurrence unadjusted and

adjusted for age, comorbidity, deprivation, tumour factors (size, lymph node status, grade) and other

treatment modalities (surgery, radiotherapy and chemotherapy)

Unadjusted Adjusted

HR 95% C.I. p HR 95% C.I. p

Adherence

≥80% 1.00 <0.001 1.00 <0.001

<80% 1.90 1.34-2.68 2.14 1.46-3.14

Ethnicity


Māori 1.27 0.80-1.99 0.309 0.99 0.61-1.63 0.979

Other 0.79 0.19-3.19 0.738 0.86 0.20-3.61 0.835

Pacific 1.15 0.36-3.61 0.817 0.42 0.12-1.40 0.158

Results

197

Figure 28 shows Kaplan-Meier survival curves demonstrating crude 5-year breast cancer

specific survival rates of 93.3% and 89.5% for women with high and sub-optimal adherence,

respectively (p=0.032). Five year disease free survival rates were 86.8% for high and 77.2%

for sub-optimally adherent women (p=0.001).

Figure 28: Kaplan-Meier survival curves for 5-year breast cancer specific and disease free survival by

adherence to adjuvant endocrine therapy in Waikato, New Zealand 2005-2011

Discussion:

From this population-based cohort study we report that Indigenous Māori and Pacific women

do have significantly higher rates of sub-optimal adherence to adjuvant endocrine therapy

compared with NZ European women. This is important especially in light of our finding that

risk of death and recurrence from breast cancer were significantly higher among women with

sub-optimal adherence. This suggests that sub-optimal adherence to endocrine therapy may be

a contributing factor to breast cancer mortality inequity between Māori and NZ European

women, although this study was not able to prove it due to limitation of numbers. To our

knowledge this is the first New Zealand study to investigate the impact of sub-optimal

adherence to adjuvant endocrine therapy as a contributor for ethnic inequities in breast cancer

survival.


198

Overall, the rate of optimum level of adherence to endocrine therapy seen in our study was

comparable to other retrospective registry based studies (217, 220, 344), although much higher

adherence rates are observed in endocrine therapy clinical trials (348). The rate of sub-optimal

adherence was significantly higher among Māori women and included more than one third of

Māori women. Even after adjusting for covariates, the odds of sub-optimal adherence for

Māori women was more than 50% higher compared to NZ European women. Failure of these

adjustments to adequately explain the observed lower adherence among Māori is most likely

due to the impact of confounders such as barriers to accessing healthcare and health literacy

(97), which were unmeasured in the present study.

Majority of women on endocrine therapy depend on general practitioners for follow up and

regular endocrine therapy prescriptions, in between annual follow ups provided by specialist

breast care clinics. Māori are known to experience more barriers and hence less access to

primary health care providers compared with NZ European patients (97). This is an important

issue since barriers that interfere with primary care may prevent general practitioner visits,

which impact on continuation of endocrine therapy. In addition, although endocrine

medications are fully funded by the government, women are required to pay a consultation or

prescription fee of NZ $15 to 50 which comes on top of travel costs, time off work and cost of

care for dependents (349). Although was not statistically significant, higher rates of

suboptimal adherence were seen among women of higher deprivation groups, and this further

supports the association between cost affordability and adherence. Māori are more likely to be

socioeconomically deprived and live in rural areas with less access to transport compared with

NZ Europeans (34). As a result, Māori women are more likely to skip general practitioner

visits which are a likely major contributor for sub-optimal adherence to endocrine therapy

(350).

Good communication, regular advice and a good physician-patient relationship improve

patient understanding and help maintain an optimum adherence with many medications

including endocrine therapy (351). A good health literacy enables a patient to process and

understand health information and promote better health decision making (103). Three out of

four Māori females have poor health literacy skills, which is approximately 50% higher than

NZ European women (106). Improving health literacy among women with breast cancer has

the potential to improve adherence to endocrine therapy as well as to increase uptake and

adherence with other adjuvant therapies, including chemotherapy and radiotherapy (352).

Whilst low health literacy is a greater issue for Māori women and hence Māori are more likely

Results

199

to benefit from any initiative aimed at improving health literacy, substantial proportions of

women of other ethnicities may also benefit due to the widespread nature of both low

adherence and low health literacy.

A limitation of our study design was the assumption that all prescribed medications were

actually consumed by the patient. Despite that, this design has been shown to provide better

estimates of adherence compared to other designs such as patient surveys or direct patient

observation and has been validated by several previous studies (217, 220, 344). Although we

managed to identify several associated factors for lower adherence, we were unable to identify

specific underlying causes for lower adherence, as reasons for sub-optimal adherence were not

available from our database. Moreover, the relatively short follow up of our study may have

underestimated the actual survival benefit of good adherence, as the benefits of endocrine

therapy are known to extend well beyond 10 years (49).

In conclusion, this study demonstrates that poor adherence to endocrine therapy is a significant

factor for higher breast cancer mortality and recurrence, and may be a contributing factor

towards breast cancer mortality inequity between Indigenous Māori and European women in

New Zealand. Improving patient understanding of benefits of adjuvant endocrine therapy

through better health literacy together with removal of existing barriers to access health care,

especially for Māori women, need to be considered as possible avenues to improve adherence.

These measures have the potential to improve adherence to care, not only for Indigenous

Māori, but for all women with breast cancer in New Zealand.


200

Results

201

5.9. Are there ethnic differences in the quality of surgical care provided for

breast cancer?

Preface:

This chapter contains an abbreviated version of a manuscript Published in ANZ Journal of

Surgery

Authors: Seneviratne S, Scott N, Lawrenson R, Campbell I.

Title: Ethnic differences in surgical treatment of breast cancer in New Zealand

Journal: ANZ Journal of Surgery

Impact factor: 1.12


DOI: 10.1111/ans.13011

Journal’s aims and scope: ANZ Journal of Surgery, established more than 70 years, is

the leading surgical journal published in Australia, New Zealand and the South-East

Asian region. The Journal is dedicated to the promotion of outstanding surgical

practice and research of contemporary and international interest. ANZ Journal of

Surgery publishes high-quality papers related to clinical practice and/or research in all

fields of surgery and its related disciplines.


202

Abstract:

Background:

Indigenous Māori are known to experience inferior quality cancer care compared with non-

Indigenous Europeans in New Zealand. However, limited data are available on

ethnic/socioeconomic differences in surgical treatment of breast cancer, or reasons for such

variations within the local context. We investigated ethnic/socioeconomic differences in rates

of mastectomy, sentinel node biopsy (SNB), post-mastectomy breast reconstruction and

definitive local therapy for breast cancer in New Zealand.

Methods:

A retrospective review of prospective data in the Waikato Breast Cancer Register for women

diagnosed during 1999-2012 was performed. Differences in rates of mastectomy (for stage

I/II, T1/T2 cancers), SNB (for stage I/II, T1/T2, cN0 cancers), post-mastectomy breast

reconstruction (for non-metastatic cancers in women <70 years) and definitive local therapy

(for stage I/II cancers) were analysed in univariate and multivariate regression models,

adjusting for covariates.

Results:

Significantly lower mastectomy and higher reconstruction rates were associated with younger

age, private compared with public hospital care and screen compared with non-screen

detection. Compared with NZ Europeans, Māori (41% vs. 33%, p=0.025) were significantly

more likely to undergo mastectomy for cancers which were potentially amenable for breast

conserving surgery, but were significantly less likely to undergo post-mastectomy breast

reconstruction (12% vs. 35%, p<0.001). No significant ethnic or socioeconomic differences

were observed in rates of SNB or definitive local therapy.

Conclusions:

This study has demonstrated lower rates of breast conserving surgery and reconstructions in

Māori compared with NZ European women, and highlight the need for future research to focus

on understanding the reasons behind these findings.

Results

203

Background:

In 1990, a Consensus Statement from the US National Institute of Health recommended that

either breast conserving surgery (BCS) followed by whole breast irradiation or total

mastectomy as local therapies of equal oncological efficacy for early stage breast cancer (353).

The efficacy and safety of sentinel lymph node biopsy (SNB) based management for clinically

node negative early breast cancer have been well established for more than a decade, and were

absolutely confirmed with the publication of the large NSABP randomised trial in 2010 (354).

SNB has been formally recommended in Australia and New Zealand for women with unifocal

breast cancers less than 3cm in size since 2008 (355) and has gained wide acceptance as the

standard of care for those women (356). A majority of breast cancers nowadays are diagnosed

in early stage, and hence, are suitable for BCS and/or SNB based management, resulting not

only in better cosmetic outcomes, but also in lower physical and psychological morbidity for a

majority of women (357, 358). Even for the minority of women requiring mastectomy due to

oncological reasons, advances and wider availability of cosmetic and reconstructive surgery

have seen a steady increase in rates of post-mastectomy breast reconstructions (359).

Many non-tumour related factors including age, comorbidity, patient/surgeon preference, and

availability and access to healthcare services have been shown to influence the decision on

type of surgical treatment for breast cancer (360-363). Many of these factors also contribute to

ethnic, socioeconomic and geographic variations in quality and type of surgical care, which

are well documented from many countries (361, 364-366). These variations include lower

rates of BCS, SNB, post-mastectomy breast reconstruction and definitive local therapy among

women of minority/Indigenous ethnicity, lower socioeconomic status and rural residency

(364-366).

At present limited data are available on quality or types of surgical treatment received by

women with breast cancer in New Zealand (197) or possible ethnic differences in such

treatment. We investigated differences in rates of BCS, SNB, post-mastectomy breast

reconstruction and definitive local therapy for breast cancer by ethnicity among a cohort of

women with invasive breast cancer in New Zealand. We also investigated tumour and socio-

demographic factors associated with these differences, and time trends in disparities in

surgical care between Māori and NZ European women.


204

Methods:

Study population:

All women with newly diagnosed primary invasive breast cancer between 1999 and 2012 were

identified (N=2848) from the WBCR.

Eligibility criteria:

Different inclusion criteria were used to identify women eligible for separate analyses.

1. Breast conserving surgery (BCS) versus mastectomy - all women with early stage

(stages I & II), T1 or T2 tumours undergoing a primary surgical intervention for breast

cancer (n=2140).

2. SNB versus non-SNB based management of the axilla - women with early stage (stage

I and II) T1 or T2 tumours with clinically node negative axillae (cN0), undergoing an

axillary surgical intervention (n=1910).

3. Post-mastectomy breast reconstruction (immediate or delayed) versus no

reconstruction - all women younger than 70 years with non-metastatic invasive breast

cancer undergoing mastectomy (primary or following failed BCS) as surgical

intervention for the ipsilateral breast (n=888).

4. Completed definitive local therapy (i.e. BCS with radiotherapy or mastectomy with or

without radiotherapy) versus non-completed definitive local therapy - all women with

stage I & II breast cancer (n=2245).

Outcome variables:

Women undergoing simple/total mastectomy, radical/modified radical mastectomy and

skin/nipple sparing mastectomy were considered to have received a mastectomy. Any

operation that was less than a mastectomy (i.e. excision biopsy, lumpectomy, wide local

excision, partial mastectomy, sector resection or quadrantectomy) was considered as a BCS.

SNB based management of the axilla was defined as all instances where a radio-isotope or a

blue dye or both were used to identify first axillary lymph node/s draining the ipsilateral breast

or the tumour/s. All other situations where the axilla was treated surgically (i.e. primary

axillary lymph node dissection or axillary node sampling) were considered as non-SNB based

management. Any woman undergoing immediate or delayed reconstruction of the ipsilateral

breast after a total mastectomy using autologous tissue, implants or both, was considered to

Results

205

have undergone a breast reconstruction. Definitive local therapy was defined as undergoing

either BCS followed by whole breast radiation or any form of mastectomy as defined above.


Chi squared (χ²) tests for trend were used to test for univariate differences, and multivariable

logistic regression models were used to test for multivariate differences in distribution of

factors associated with BCS versus mastectomy, SNB versus non-SNB based management of

the axilla, post-mastectomy breast reconstruction versus non-reconstruction and completed

versus incomplete definitive local therapy. As some of the variables included of missing data,

analyses were repeated using only cases with complete data for all variables. The results were

almost identical to the full dataset, and are not presented in this report. Imputation of missing

values was not undertaken due to the similarity of these results.

Results:

Breast Conserving Surgery (BCS) versus Mastectomy:

Of a total of 2140 women with early stage T1 and T2 tumours undergoing a primary surgical

treatment for the ipsilateral breast, 751 (35.1%) underwent primary mastectomy and 1389

(64.9%) underwent primary BCS. Māori compared with NZ European ethnicity (OR=1.45,

95% CI 1.07-1.95), a distance of >50km from surgical treatment facility (OR=1.48, 95% CI

1.13-1.93), age above 70 years (OR=1.66, 95% CI 1.19-2.34), non-screen detection (OR=1.79,

95 % CI 1.42-2.28), T2 primary tumour (OR=2.30, 95% CI 1.86-2.84), lobular histology

(OR=1.64, 95% CI 1.19-2.28), multi-focality (OR=2.75, 95% CI 2.11-3.58), treatment in a

public hospital (OR=1.26, 95% CI 1.01-1.60) and a comorbidity score ≥1 (OR=1.48, 95% CI

1.12-1.93) were significantly associated with mastectomy rather than BCS as primary surgical

treatment in the multivariable logistic regression model (Table 1). Peak rate of primary

mastectomy (40.7%) was observed during 2003-2006, which since has declined gradually to

just under 30% during 2010-2012. As Māori women had a higher proportion of T2 cancers

compared with NZ European women (42.9% vs. 35.1%), rates of mastectomy were analysed

separately for T1 and T2 cancers between these two groups. Higher rates of mastectomy were

observed in Māori women for both T1 (31.7% vs. 25.2%) and T2 (55.3% vs. 49.7%) tumours.


206

Table 36: Characteristics associated with mastectomy versus breast conserving surgery for women

with T1 & T2 breast cancers undergoing surgery in the Waikato 1999-2012

Characteristic

Total

(N=2140)

n (%)

Mastectomy

(%)

Unadjusted Adjusted


Ethnicity

NZ European 1774 (82.9) 599 (33.8) Ref Ref

Māori 287 (13.4) 120 (41.8) 1.41 1.09-1.82 0.008 1.45 1.07-1.95 0.015

Pacific 30 (1.4) 12 (40.0) 1.31 0.63-2.73 0.476 1.19 0.54-2.64 0.660

Other 49 (2.3) 20 (40.8) 1.35 0.76-2.41 0.306 1.67 0.88-3.15 0.115

Age (yrs.)

<40 85 (4.0) 40 (47.1) 1.64 1.02-2.63 1.50 0.91-2.48

40-59 996 (46.6) 322 (32.3) Ref <0.001 Ref <0.001

60-79 897 (41.9) 303 (33.8) 0.98 0.78-1.14 1.03 0.81-1.32

80+ 162 (7.6) 86 (53.1) 2.09 1.44-3.02 1.72 1.14-2.67

Deprivation

Dep 1-2 226 (10.6) 81 (35.8) Ref 0.733 Ref 0.811

Dep 3-4 232 (10.8) 73 (31.5) 0.82 0.56-1.21 0.80 0.52-1.22

Dep 5-6 544 (25.4) 187 (34.4) 0.94 0.68-1.30 0.83 0.58-1.19

Dep 7-8 608 (28.4) 217 (35.7) 0.99 0.72-1.37 0.86 0.59-1.26

Dep 9-10 530 (24.8) 193 (36.4) 1.03 0.74-1.42 0.80 0.54-1.18

Distance

<10km 661 (30.9) 214 (32.4) Ref 0.163 Ref 0.045

10-50km 843 (39.4) 294 (34.9) 1.12 0.90-1.39 1.14 0.86-1.55

>50km 636 (29.7) 210 (39.8) 1.31 1.04-1.67 1.48 1.13-1.93

Screening status

Screen 935 (43.7) 213 (22.8) Ref <0.001 Ref <0.001

Non-screen 1205 (56.3) 538 (44.6) 2.73 2.26-3.31 1.79 1.42-2.28

Hospital type

Private 672 (31.4) 221 (32.9) Ref 0.148 Ref 0.048

Public 1468 (68.6) 530 (36.1) 1.15 0.95-1.40 1.26 1.01-1.60

Results

207

Diagnosis year

1999-2002 473 (22.1) 165 (34.9) Ref <0.001 Ref <0.001

2003-2006 649 (30.3) 264 (40.7) 1.28 1.00-1.64 1.54 1.14-2.06

2007-2009 485 (22.7) 168 (34.6) 0.99 0.76-1.29 1.07 0.75-1.52

2010-2012 533 (24.9) 154 (28.9) 0.76 0.58-0.99 0.79 0.56-1.13

Grade

Grade I 584 (27.3) 157 (26.9) Ref <0.001 Ref 0.831

Grade II 1100 (51.4) 407 (37.0) 1.60 1.28-1.99 1.11 0.86-1.43

Grade III 419 (19.6) 174 (41.5) 1.93 1.48-2.52 1.15 0.82-1.61

Unknown 37 (1.7) 13 (35.1)

ER/PR status

ER/PR + 1817 (84.9) 617 (34.0) Ref 0.012 Ref 0.089

ER & PR - 302 (14.1) 125 (41.4) 1.37 1.07-1.76 1.31 0.94-1.78

Unknown 21 (1.0) 9 (42.9)

HER-2 status

Negative 1171 (54.7) 382 (50.9) Ref 0.004 0.136

Equivocal 135 (6.3) 54 (7.2) 1.38 0.96-1.98 1.25 0.82-1.95

Positive 244 (11.4) 108 (14.4) 1.64 1.24-2.17 1.41 1.03-1.91

Unknown 590 (27.6) 207 (27.6)

Histology type

Ductal 1772 (82.8) 603 (34.0) Ref Ref 0.005

Lobular 216 (10.1) 102 (47.2) 1.73 1.30-2.31 0.000 1.64 1.19-2.28

Other 152 (7.1) 46 (30.3) 0.84 0.59-1.21 0.346 0.82 0.55-1.21

Multi-focality

No 1813 (84.7) 580 (32.0) Ref <0.001 Ref <0.001

Yes 327 (15.3) 171 (52.3) 2.33 1.84-2.96 2.75 2.11-3.58

T stage

T1 1356 (63.3) 353 (26.0) Ref <0.001 Ref <0.001

T2 784 (36.7) 398 (50.8) 2.93 2.43-3.53 2.30 1.86-2.84

Charlson score

0 1790 (83.6) 591 (33.0) Ref <0.001 Ref 0.015

≥1 350 (16.3) 160 (45.7) 1.65 1.29-2.09 1.48 1.12-1.93


208

Of women who underwent primary mastectomy (n=751), details of decision process for

mastectomy (surgeon recommendation versus patient preference) was documented for 600

(79.9%) women. Of these, 139 (23.2%) mastectomies were due to patient preference.

Compared with NZ European women (21.5%), a higher proportion of mastectomies in Māori

(24.7%) was due to patient preference, although this difference was statistically not significant

(p=0.560). No significant differences in tumour or socio-demographic characteristics were

observed between women for whom the decision was documented compared with women for

whom it was not (Appendix 8).

SNB versus non-SNB based management of the axilla:

The rate of SNB for early stage T1 and T2 cancers has increased by four-fold over the study

period from, 21.3% during 1999-2002 to 84.7% during 2010-2012. Overall rates of SNB

between Māori and NZ European women were similar (55.1% and 55.4%, respectively,

p=0.921) and similar increasing trends in SNB rates were observed for Māori (from 7.9% to

79.3%) and NZ European women (from 23.5% to 85.9%) over the study period (Figure 29).

Multivariable regression model adjusting for socio-demographic and tumour characteristics

also did not show a significant difference in rate of SNB (OR=1.16, 95% CI 0.96-1.54)

between Māori and NZ European women (Appendix 9).

Figure 29: Trends in rate of sentinel lymph node biopsy for women with early stage (stage I & II), T1-

2, cN0 tumours undergoing an axillary surgical intervention by ethnicity

Breast reconstruction following mastectomy

Unadjusted and adjusted rates of post-mastectomy breast reconstruction for non-metastatic

breast cancer in women younger than 70 years are shown in Table 37. Overall, 263 (29.6%)

0%

20%

40%

60%

80%

100%

1999-2002 2003-2006 2007-2009 2010-2012

NZ European

Māori

Results

209

women had received a post-mastectomy reconstruction of which 237 (90.1%) were immediate

and 26 (9.9%) were delayed reconstructions. Māori compared with NZ European ethnicity

(OR=0.38, 95% CI 0.18-0.64), age above 50 years (OR=0.56, 95% CI 0.38-0.82), public

hospital care (OR=0.56, 95% CI 0.39-0.81), primary tumour larger than T2 (OR=0.52, 95%CI

0.36-0.76) and comorbidity (OR=0.46, 95% CI 0.26-0.84) were significant predictors of not

receiving a breast reconstruction, in the multivariable model. Higher body mass index (BMI)

was also associated with lower reconstruction rates, but was statistically significant only for

women with a BMI ≥35 (OR=0.37, 95% CI 0.14-0.99) in the multivariable model. The Māori

cohort who underwent mastectomy tended to be younger than their NZ European counterparts

(mean age 51.6 vs. 52.5 years, p=0.457) and compared with NZ Europeans, higher proportions

of Māori women were observed to be obese (57.7% vs. 25.3%, p<0.001), smokers (59.6% vs.

30.4%, p<0.001), belonged to the two highest deprivation quintiles (70.3% vs. 48.4%,

p<0.001) and a lower proportion (11% vs. 37.9%, p<0.001) had received surgery in the private

sector. Adjusting for these factors resulted only in a marginal change in the odds of breast

reconstruction for Māori women. The rate of reconstruction for Māori has not changed

substantially over the study period (varying between 10 and 15%), while a 50% increase in the

rate of breast reconstruction was observed for NZ European women (Figure 30).

Table 37: Characteristics associated with women undergoing major breast reconstruction following

mastectomy for breast cancer in Waikato 1999-2012 a

Characteristic

Total

(N=888)

n (%)

Reconstruction

(N=263)

n (%)

Unadjusted Adjusted


Ethnicity


Māori 172 (19.4) 21 (12.2) 0.25 0.16-0.41 <0.001 0.38 0.18-0.64 <0.001

Pacific 24 (2.7) 5 (20.8) 0.48 0.18-1.31 0.151 0.48 0.16-1.50 0.217

Other 33 (3.7) 4 (12.1) 0.25 0.09-0.73 0.011 0.16 0.05-0.49 0.001

Age (years)

<40 82 (9.2) 42 (51.2) 1.52 0.93-2.50 1.81 1.03-3.18

40-49 267 (30.1) 109 (40.8) Ref <0.001 Ref <0.001

50-59 299 (33.7) 93 (31.1) 0.65 0.46-0.92 0.56 0.38-0.82

60-69 240 (27.0) 19 (7.9) 0.12 0.07-0.21 0.11 0.06-0.18


210

Deprivation

Dep 1-2 89 (10.0) 41 (46.1) Ref 0.004 Ref 0.576

Dep 3-4 95 (10.7) 28 (29.5) 0.49 0.27-0.90 0.56 0.28-1.13

Dep 5-6 232 (26.1) 73 (31.5) 0.54 0.33-0.89 0.68 0.38-1.22

Dep 7-8 239 (26.9) 65 (27.2) 0.44 0.26-0.72 0.71 0.39-1.28

Dep 9-10 233 (26.2) 56 (24.0) 0.37 0.22-0.62 0.76 0.41-1.39

Hospital type

Private 280 (31.5) 118 (42.1) Ref <0.001 Ref 0.002

Public 608 (68.5) 145 (23.8) 0.43 0.32-0.58 0.56 0.39-0.81

Year of diagnosis

1999-2002 196 (22.1) 53 (27.0) Ref 0.290 Ref 0.035

2003-2006 310 (34.9) 88 (28.4) 1.07 0.72-1.60 1.06 0.67-1.67

2007-2009 201 (22.6) 58 (28.9) 1.09 0.71-1.70 1.66 0.99-2.76

2010-2012 181 (20.4) 64 (35.4) 1.48 0.95-2.29 1.80 1.08-2.98

T stage

T1 349 (39.3) 127 (36.4) Ref 0.004 Ref 0.001

T2 416 (46.8) 109 (26.2) 0.62 0.46-0.85 0.52 0.36-0.76

T3 75 (8.4) 18 (24.0) 0.55 0.31-0.98 0.48 0.25-0.92

T4 47 (5.3) 9 (19.1) 0.41 0.19-0.88 0.29 0.13-0.68

Charlson score

0 761 (85.7) 250 (32.9) Ref <0.001 Ref 0.018

≥1 127 (14.3) 13 (10.2) 0.25 0.14-0.45 0.46 0.26-0.84

BMI category

<20 39 (4.4) 14 (35.9) Ref <0.001 Ref 0.058

20-30 471 (53) 165 (35.0) 0.98 0.48-1.98 1.07 0.48-2.42

30-35 141 (15.9) 40 (28.4) 0.75 0.33-1.50 1.13 0.48-2.68

>35 107 (12.0) 13 (12.1) 0.25 0.10-0.59 0.37 0.14-0.99

Unknown 130 (14.6) 31 (28.3) 0.56 0.26-1.21 0.86 0.36-2.08

Smoking status

Never smoked 536 (60.4) 170(31.7) Ref 0.045 Ref 0.312

Smoker 295 (33.2) 77 (26.1) 0.70 0.49-0.99 0.80 0.53-1.23

Unknown 57 (6.4) 16 (28.1) 0.84 0.46-1.54 0.89 0.44-1.44

a non-metastatic breast cancer among women <70 years

Results

211

Figure 30: Trends in the rates of post-mastectomy breast reconstruction by ethnicity

Definitive local therapy for early stage (stage I & II) breast cancer

Overall, 89.5% of NZ European and 91.1% of Māori women had completed definitive local

therapy (p=0.324) over the study period. Age older than 70 years (OR=0.46, 95% CI 0.26-

0.82), higher tumour grade (OR=1.56, 95% CI 1.06-2.24) and comorbidity (OR=0.55, 95% CI

0.38-0.80) were significantly associated with non-completed definitive local therapy in the

multivariable logistic regression model (Table 38).

Table 38: Characteristics associated with women completing definitive local therapy for early (stage I

& II) breast cancer in Waikato 1999-2012

Characteristic Total

(N=2245)

Definitive

local therapy

completed

Unadjusted Adjusted

n (%) n (%) OR 95% CI p OR 95% CI p

Ethnicity


Māori 305 (13.6) 278 (91.1) 1.21 0.80-1.85 0.367 0.88 0.54-1.44 0.608

Pacific 32 (1.4) 28 (87.5) 0.83 0.29-2.38 0.722 0.76 0.19-3.12 0.708

Other 50 (2.2) 48 (96.0) 2.83 0.68-11.7 0.152 1.23 0.29-5.33 0.778

0%

10%

20%

30%

40%

50%

1999-2002 2003-2006 2007-2009 2010-2012

NZ European

Māori


212

Age (yrs.)

<40 88 (3.9) 88 (100) - - - -

40-49 410 (18.3) 388 (94.6) Ref <0.001 Ref <0.001

50-59 603 (26.8) 569 (94.4) 0.95 0.55-1.65 1.01 0.57-1.80

60-69 583 (25.9) 540 (92.6) 0.71 0.42-1.21 0.82 0.47-1.44

70-79 343 (15.3) 299 (87.2) 0.39 0.23-0.66 0.46 0.26-0.82

80+ 218 (9.7) 132 (60.6) 0.09 0.05-0.15 0.16 0.09-0.29

Deprivation

Dep 1-2 231 (10.3) 217 (93.9) Ref 0.114 Ref 0.172

Dep 3-4 245 (10.9) 213 (86.9) 0.43 0.22-0.83 0.39 0.18-0.86

Dep 5-6 572 (25.5) 519 (90.7) 0.63 0.34-1.16 0.57 0.27-1.18

Dep 7-8 641 (28.5) 573 (89.4) 0.54 0.30-0.99 0.57 0.28-1.17

Dep 9-10 556 (24.8) 494 (88.8) 0.51 0.28-0.94 0.47 0.22-0.97

Diagnosis year

1999-2002 494 (22.0) 432 (87.4) Ref 0.144 Ref 0.210

2003-2006 679 (30.2) 615 (90.6) 1.38 0.95-1.99 1.49 0.95-2.32

2007-2009 513 (22.8) 470 (91.6) 1.57 1.04-2.36 1.37 0.84-2.21

2010-2012 559 (24.9) 499 (89.3) 1.19 0.82-1.74 1.03 0.66-1.61

Grade

Grade I 589 (26.2) 529 (89.8) Ref <0.001 Ref <0.001

Grade II 1124 (50.1) 1041 (92.6) 1.42 1.01-2.02 1.54 1.06-2.24

Grade III 436 (19.4) 411 (94.3) 1.86 1.15-3.03 1.79 1.01-3.17

Unknown 96 (4.3) 35 (36.5)

ER/PR status

ER &/or PR + 1874 (83.5) 1686 (90.0) Ref 0.048 Ref 0.705

ER & PR - 335 (14.9) 313 (93.4) 1.59 1.01-2.51 0.90 0.52-1.05

Unknown 36 (1.6) 17 (47.2)

T stage

T1 1378 (61.4) 1257 (91.2) Ref 0.004 Ref 0.424

T2 829 (36.9) 733 (88.4) 0.79 0.59-1.06 1.08 0.76-1.53

T3 38 (1.7) 28 (73.7) 0.29 0.13-0.62 0.56 0.21-1.50

Charlson score

0 1838 (81.9) 1707 (92.9) Ref Ref <0.001

≥1 407 (18.1) 309 (75.9) 0.30 0.22-0.41 0.55 0.38-0.80

Results

213

Discussion:

This study has shown that Indigenous Māori women were significantly more likely to undergo

primary mastectomy for breast cancers which were potentially amenable for BCS, but were

significantly less likely to receive a post-mastectomy breast reconstruction. Rates of SNB and

definitive local therapy did not differ significantly between Māori and NZ European women.

Overall, the rates of BCS, SNB, breast reconstruction and definitive local therapy were

acceptable and comparable to rates reported from countries with similar health care systems

(367, 368).

Mastectomy versus BCS:

Patients’ preference for mastectomy over BCS appears to have been a major contributor, while

surgeon preference may also have contributed for higher rates of mastectomy observed in

Māori compared with NZ European women. Although details of mastectomy decision process

was available for only 80% of women, this explanation is likely to be valid as no significant

differences in tumour or socio-demographic characteristics were observed between women for

whom details of decision process was available and the rest.

As expected, tumour size and multi-focality influenced mastectomy, but no significant

associations were observed between tumour biological characteristics which are known to be

associated with higher risks of local or systemic failure, and the rate of mastectomy. The

association between aggressive tumour characteristics and higher mastectomy rate is well

documented (361), and is due to the belief among some surgeons that BCS is an oncologically

inferior operation to mastectomy, especially for more aggressive cancers (361). The absence

of this observation indicates widespread acceptance of BCS for all early breast cancers among

surgeons in the region, and adherence to treatment guidelines (44).

Several factors are known to influence a woman’s decision towards mastectomy over BCS, for

a cancer that is technically suitable for BCS. These include; the belief that mastectomy is

“safer”, to avoid radiotherapy due to fear of its long term complications or due to difficulties

in accessing radiotherapy services (such as living some distance from a treatment centre) and

to avoid potential re-operations (360, 369-372). Women of low socioeconomic backgrounds

and rural women have been documented to have significantly higher rates of mastectomy

compared with socioeconomically affluent and urban women, respectively as they are more

likely to experience difficulties in accessing radiotherapy services (80, 361, 373). Consistent

with this, a significantly higher mastectomy rate was observed among rural compared with


214

urban dwelling women. However, no similar association was observed between

socioeconomic deprivation and rates of mastectomy. Lack of this association in the present

study might have been influenced at least to some extent, by the use of an area based

deprivation system to measure socioeconomic status. A significantly lower rate of mastectomy

among women treated in private versus public hospitals was also observed. This, though

indirectly, supports the association between higher mastectomy rates with lower


Post-mastectomy breast reconstruction:

A significantly higher rate of breast reconstruction was observed among NZ European women

compared with Māori women. However, whether this difference was contributed by patient or

surgeon decision process could not be evaluated as this information was not available from the

WBCR database. Regardless, we observed several factors which appear to have contributed to

this disparity including higher socioeconomic deprivation, obesity, smoking and lower

likelihood of private sector treatment in Māori compared with NZ European women. The

difference in rates of reconstruction between public and private seems to have had a major

impact on difference in rates of reconstruction between Māori and NZ European as only 10%

of Māori received treatment in private sector compared with 38% of NZ European women.

Higher socioeconomic deprivation in Māori compared with NZ European women seems to

have contributed to this disparity as women of more affluent backgrounds were more likely to

receive treatment from the private sector. Seeking reconstruction from the private sector might

also have been influenced by lack of availability and/or longer wait lists for breast

reconstruction in public sector. It is likely that at least for some women, who could not afford

private sector reconstructions, these delays would have prompted to forgo reconstruction.

Lower uptakes of breast reconstruction in minority ethnic women have been documented in

the USA, which were contributed by several reasons including cultural, educational and

financial factors (366, 374). It is unclear whether cultural factors and possible differences in

attitudes towards body image have contributed to differences in rates of BCS and

reconstruction between Māori and NZ European women.

SNB and definitive local therapy:

There were no differences in rates of SNB and definitive local therapy by either ethnicity or

socioeconomic deprivation. Rates of SNB have steadily increased over the study period as

SNB gained recognition as standard of care for women with clinically node negative early

Results

215

breast cancer. However, the safety of SNB for early breast cancers >3cm in diameter or for

multifocal cancers has not been proven by randomised trial yet, and women with likely

involved nodes on imaging or clinically, are not offered SNB. These are the likely reasons for

the SNB rate plateauing at 80% during 2010-2012.

Overall, from this study, it appears that interventions which are associated with cancer

outcomes (i.e. definitive local therapy) and morbidity (i.e. SNB) have been provided and taken

up by women of diverse ethnic and socioeconomic groups at almost similar rates, while major

differences were observed for interventions that improve cosmetic outcomes (i.e. BCS and

reconstruction). Breast cancer surgery is no longer considered an oncological intervention

aimed only at cancer cure, but also as an intervention aimed at providing best quality of life

for a woman, by preserving a near normal breast or by creating a new breast through

reconstruction. Therefore, it is important that all women are provided with the option of

reconstruction, unless it is contraindicated due to patient or tumour factors. However this

needs to be provided in a manner that is acceptable to women of different ethnic, cultural and

socioeconomic backgrounds, and perhaps more importantly in a way that it would not

compromise long term cancer outcomes, for instance by adding longer treatment delays.

There were a few limitations in this study. Despite the WBCR being a comprehensive database

that captures all surgical treatment in detail, some of the delayed reconstructions, especially

the ones performed outside the Waikato region may have been omitted. However such

numbers are expected to be minimal and the undercounting resulted from this bias is unlikely

to influence the final results of this study.

In conclusion, a majority of study women were observed to have received high quality surgical

care for breast cancer on par with accepted guidelines. However, significant disparities in

quality of surgical care for breast cancer by ethnicity and geographic location were observed

with Māori and rural women receiving fewer BCS and reconstructions compared with NZ

European and urban women, respectively. Many socio-demographic and healthcare services

related factors likely to have contributed to these observed differences. However, exact

mechanisms and their contributions towards lower rates of BCS and breast reconstructions in

Māori compared with NZ European women are unclear at present. Future research focussing

on understanding underlying mechanisms for these differences is needed which may help

develop measures to reduce disparities.


216

Results

217

5.10. How things add up: quantitative impact of factors on ethnic inequity

Preface:

This chapter contains an abbreviated version of a manuscript submitted for publication in

Cancer Causes and Control

Authors: Seneviratne S, Campbell I, Scott N, Shirley R, Peni T, Lawrenson R.

Title: Ethnic differences in breast cancer survival in New Zealand:

Contributions of differences in screening, treatment, tumour biology,

demographics and comorbidities

Journal: Cancer Causes & Control

Impact factor: 2.96

Journal’s aims and scope: Cancer Causes & Control is an international refereed

journal that both reports and stimulates new avenues of investigation into the causes,

control, and subsequent prevention of cancer. Its multidisciplinary and multinational

approach draws together information published in a diverse range of journals.

Coverage of the journal extends to variation in cancer distribution within and between

populations; factors associated with cancer risk; preventive and therapeutic

interventions on a population scale; economic, demographic, and health-policy

implications of cancer; and related methodological issues.


218

Abstract:

Introduction:

Underlying reasons for ethnic inequities in breast cancer outcomes are complex and poorly

understood. We investigated the breast cancer survival inequity between Indigenous Māori

and European women, and quantified relative contributions of patient, tumour and healthcare

system factors towards this inequity.

Methods:

All women with newly diagnosed invasive breast cancer between 1999 and 2012 were

identified from the WBCR. Cancer specific survival between Māori and NZ European women

was compared using Kaplan-Meier survival curves while contributions of different factors

towards the survival disparity were quantified with Cox proportional hazard modelling.

Results:

Of the total of 2791 women included in this study, 2260 (80.1%) were NZ European and 419

(15%) were Māori. Compared with NZ European women, Māori had a significantly higher age

adjusted cancer specific mortality (HR =2.02, 95% CI, 1.59-2.58) with significantly lower 5-

year (86.8% vs. 76.1%, p<0.001) and 10-year (79.9% vs. 66.9%, p<0.001%) crude cancer-

specific survival rates. Stage at diagnosis explained approximately 40% while screening,

treatment and patient factors (i.e. comorbidity, obesity and smoking) contributed by

approximately 15% each towards the survival disparity. The final model accounted for almost

all of the cancer survival disparity between Māori and NZ European women (HR=1.07, 95%

CI, 0.80-1.44).

Conclusions:

Māori women experience an age-adjusted risk of death from breast cancer, which is more than

twice that for NZ European women. Lower screening coverage, delay in diagnosis, inferior

quality of treatment and greater patient comorbidity appear to be important factors

contributing to survival disparity between Māori and NZ European women.

Results

219

Introduction:

Ethnic disparities in cancer survivals are well documented for a range of cancers among many

different ethnic populations.

The underlying factors for ethnic disparities in cancer survival are complex, poorly understood

and vary widely among different countries and populations (96, 375). Causes of inequities in

cancer survival between ethnic and socioeconomic groups can be categorized broadly into

healthcare system, patient, and tumour related factors. Patient level factors include access to

the determinants of health such as income, healthy housing, education and a healthy

environment. Inequities in access to these determinants of health cause inequities in patient

level factors such as comorbidities, smoking and obesity. Only a few studies to date have

reported on the impact of comorbidities, smoking and obesity on breast cancer survival

disparity. Many studies on cancer survival rely on routinely collected data sets such as

national cancer registries, and data on these variables are not available from routine datasets.

(146). The impact of tumour biological characteristics has been studied in great detail,

especially in the USA, where Black African women are known to have breast cancers with

biological characteristics associated with worse outcomes (166). Healthcare system

contributes to survival disparity through differences in barriers to access care (135) and once

gained access, through institutional factors including discrimination, lower use and inferior

quality cancer treatment provided for Indigenous/ethnic minority patients (307, 316).

This study was aimed at investigating the breast cancer survival disparity between Māori and

NZ Europeans in a regional cohort of New Zealand women, and to quantify the relative

contributions of patient, tumour and healthcare system factors towards the survival disparity.

Methods:

Data sources:

Data were obtained from the WBCR and the WBCR data were linked with the National

Mortality Collection, the National Breast Cancer Screening Database and the National

Minimum Dataset (NMDS) using National Health Index number.


220

Study population:

A total of 2791 women with newly diagnosed breast cancer (2848 cancers) over a 14-year

period from 01/01/1999 to 31/12/2012 were identified from the WBCR. For women with more

than one episode of invasive cancer, data from the first episode were included for analysis.

Study covariates:

Breast cancer treatment

Breast cancer treatment was broadly considered under local and systemic modalities.

Definitive local therapy was defined as receipt of breast conserving surgery (BCS) followed

by radiation therapy or receipt of mastectomy that was performed with curative intent. BCS

without radiation, palliative mastectomy or no surgical treatment was considered as

incomplete definitive local therapy. Receipt of chemotherapy, radiotherapy and hormonal

therapy were considered under systemic breast cancer treatment. Delays in initiating treatment

beyond proven clinically significant thresholds were considered as treatment delays. These

included a 60-day threshold from diagnosis to first surgical intervention (206), a 60-day

threshold from first surgery to chemotherapy (207, 208) and for radiotherapy, a 90-day

threshold from first surgery, where chemotherapy was not given, or from the date of

completing chemotherapy, where chemotherapy was given (209).

Outcome variables:

Date and cause of death for all deceased women (censored at 31/12/2013) were identified from

the WBCR and the National Mortality Collection. Follow up duration was calculated from the

date of diagnosis to date of death, or to the date of the last known follow up (censored at

31/12/2013).


Multivariable Cox proportional hazard models were used to calculate breast cancer specific

mortality hazard ratios with 95% confidence intervals for Māori compared with NZ European

women, stratifying into tumour stage at diagnosis (I/II and III/IV). Kaplan-Meier survival

curves were created for crude breast cancer specific mortality for Māori and NZ European

women. Due to small numbers, Pacific and Other ethnic group women were excluded from

these analyses.

Results

221

The initial base model calculated hazard ratios for breast cancer specific mortality controlling

only for age and year of diagnosis. Additional variables were introduced sequentially, starting

with breast cancer screening followed by cancer stage at diagnosis, biological characteristics,

cancer treatment, comorbidities and healthcare access factors. We also performed a separate

model with a different sequential introduction as we felt that healthcare access factors could

influence stage at diagnosis, as well as cancer treatment. For the second model, healthcare

access factors were introduced first, followed by other factors, in same sequence as the first

model.

As some of the variables included high numbers of missing data, survival analysis was

repeated using only cases with complete data for all variables. Results were almost similar to

those obtained from the Kaplan-Meier and Cox proportional hazards regression models, and

these data are not presented in this report. Imputation of missing values was not undertaken

due to the similarity of these results.

Results

Of the 2791 women included in this study, 419 (15%) were Māori, 2260 (80.9%) were NZ

European, 51 (1.8%) were Pacific and 61 (2.2%) were of Other ethnicity. A minimum follow

up of five years or up to death was available for 255 (60.8%) Māori and 1527 (67.6%) NZ

European women. Median follow up was 43 months for Māori women (mean 54.5 months,

SD=41) and 61 months for NZ European women (mean 68.9 months, SD=45). A total of 131

(31.3%) deaths were observed for Māori and 544 (24.1%) for NZ European women of which

87 (66.4%) for Māori and 312 (57.4%) for NZ European women were due to breast cancer.

The Māori cohort was significantly younger than the NZ European cohort (Table 39), in

keeping with the younger age structure of Māori population in New Zealand (31). Māori were

significantly more likely to be diagnosed with more advanced breast cancer, and were around

two and a half times more likely to be diagnosed with metastatic disease than NZ European

women (11% vs. 4.6%). Māori women had a significantly higher prevalence of medical

comorbidities (27.9% vs. 17.3%, p<0.001), obesity (52% vs. 28%, p<0.001), smoking (59%

vs. 25%, p<0.001) and were significantly more likely to live in more deprived areas (p<0.001)

compared with NZ European women. Māori women were significantly less likely to be

diagnosed through mammographic screening (29.8% vs. 36.9%, p<0.001) and were less than a


222

third as likely as NZ European women to have had surgical treatment from a private facility

(9.3% vs. 31.9%, p<0.001).

Māori women had generally higher graded cancers; fewer grade I and more grade II tumours,

and a higher rate of HER-2 positive cancers (21.9% vs. 17.2%, p<0.001) compared with NZ

European women. Māori were significantly more likely to undergo mastectomy than breast

conserving surgery and were significantly less likely to complete definitive local therapy

compared with NZ European women (79.8% vs. 83.5%, p=0.021). Māori women were

significantly more likely to have received chemotherapy (36.8% vs. 31.6%, p=0.038), but

were less likely, albeit non-significantly, to have received adjuvant endocrine therapy (67.1%

vs. 71.6%, p=0.058).

Table 39: Patient, tumour treatment and healthcare access characteristics of the study cohort by Māori

and NZ European ethnicity

Characteristic Total (N=2791)

n (%)

NZ European

(N=2260)

n (%)

Māori

(N=419)

n (%)

p

Age (mean +/- SD) 60.5 +/-13.8 61.4 +/-13.9 55.6 +/-12.2 <0.001

Diagnosis year <0.001

1999-2002 597 21.4 502 22.2 73 17.4

2003-2006 853 30.6 727 32.2 99 23.6

2007-2009 652 23.4 497 22.0 117 27.9

2010-2012 689 24.7 534 23.6 130 31.0

Stage <0.001

I 1112 39.8 943 41.7 139 33.2

II 1087 38.9 880 38.9 159 37.9

III 433 15.5 332 14.7 75 17.9

IV 159 5.7 105 4.6 46 11.0

Deprivation <0.001

Dep 1-2 284 10.2 261 11.5 13 2.1

Dep 3-4 289 10.4 248 11.0 29 6.9

Dep 5-6 678 24.3 581 25.7 73 17.4

Dep 7-8 813 29.1 655 29.0 128 30.5

Dep 9-10 727 26.0 515 22.8 176 40.2

Results

223

Urban rural status 0.008

Urban 1493 53.5 1215 53.8 201 48.0

Semi-urban 786 28.2 615 27.2 145 34.6

Rural 512 18.3 430 19.0 73 17.4

Screening status 0.006

Screen detected 988 35.4 833 36.9 125 29.8

No-screen detected 1803 64.6 1427 63.1 294 70.2

Hospital type <0.001

Private 781 28.0 722 31.9 39 9.3

Public 2010 72.1 1538 68.0 380 90.7

Charlson score <0.001

0 2264 81.1 1869 82.7 302 72.1

1-2 478 17.1 357 15.8 103 24.6

3+ 49 1.8 34 1.5 14 3.3

Smoking <0.001

Non smoker 1802 69.9 1558 75.0 163 40.8

Ex-smoker 260 10.1 204 9.8 52 13.0

Current smoker 517 20.0 315 15.2 184 46.1

Unknown (212) (183) (20)

BMI <0.001

<20 113 5.6 96 6.1 13 3.9

20-25 594 29.5 501 31.6 66 19.9

25-30 650 32.3 543 34.2 79 23.8

>30 655 32.6 446 28.1 174 52.4

Unknown (780) (674) (87)

Grade 0.007

Grade I 615 23.7 529 25.1 68 17.8

Grade II 1365 52.6 1095 52.0 221 58.0

Grade III 613 23.6 482 22.9 92 24.1

Unknown (198) (154) (38)

ER/PR status 0.153

ER &/or PR + 2298 84.1 1873 84.8 335 81.3

ER & PR - 434 15.9 336 15.2 77 18.7

Unknown (59) (51) (7)


224

HER-2 <0.001

Negative 1517 73.6 1214 74.7 251 72.3

Equivocal 156 7.6 131 8.1 20 5.8

Positive 384 18.7 280 17.2 76 21.9

Unknown (734) (635) (72)

Loco-regional treatment <0.001

BCS + Radiotherapy 1257 45.0 1064 47.1 149 35.6

Mastectomy 1065 38.2 830 36.7 183 43.7

BCS without radiotherapy 228 8.2 192 8.5 30 7.2

No primary surgery 241 8.6 174 7.7 57 13.6

Chemotherapy 0.038

Yes 924 33.1 714 31.6 154 36.8

No 1867 66.9 1546 68.4 265 63.2

Endocrine therapy 0.058

Yes 1974 70.7 1619 71.6 281 67.1

No 817 29.3 641 28.4 138 32.9

Delay in surgery a 0.001

No 2220 87.1 1841 88.3 295 81.5

Yes 330 12.9 245 11.7 67 18.5

No surgery (241) (174) (57)

Delay in chemotherapy b 0.103

No 621 67.4 494 69.5 96 62.7

Yes 301 32.6 217 30.5 57 37.3

No chemotherapy (1869) (1549) (266)

Delay in radiotherapy c 0.045

No 677 68.0 592 69.4 68 60.2

Yes 319 32.0 261 30.6 45 39.8

No radiotherapy (1795) (1407) (306)

a delay in surgery longer than 60 days from date of diagnosis, b delay in adjuvant chemotherapy longer

than 60 days from date of first surgery, c delay in radiotherapy longer than 90 days from first surgery if

no chemotherapy was given or from completion of chemotherapy if chemotherapy was given

Results

225

Breast cancer survival

Māori had a significantly lower crude cancer-specific survival than NZ European women

(p<0.001) (Figure 31). The crude five and 10-year cancer-specific survival rates were 86.8%

and 79.9% respectively, for NZ European and 76.1% and 66.9% respectively, for Māori

women. An improvement in 5-year survival rates were observed for both Māori and NZ

European women over time, which was greater for Māori than NZ European women. For

instance, 5-year survival increased from 73% to 79% for Māori from 1999-2005 to 2006-2012

while an increase from 87% to 89% was observed for NZ European women over the same

time periods (Appendix 11).

Figure 31: Kaplan-Meier survival curves for breast cancer specific survival (unadjusted) for Māori and

NZ European cohorts

Hazards ratios (HR) for breast cancer mortality for selected characteristics from the

multivariable Cox proportional model are shown in Table 40. Non-screen detection compared

with screen detection, advanced cancer stage, ER/PR negativity, higher grade, not completing

definitive local therapy and higher comorbidity index were significantly associated with higher

risks of breast cancer mortality.


226

Table 40: Breast cancer specific mortality hazard ratios for selected variables from final Cox

proportional hazard model

Characteristic HR 95% CI

Ethnicity NZ European Ref

Māori 1.07 0.80-1.44

Mode of diagnosis Screen detected Ref

Non-Screen 1.44 1.05-1.96

T stage 1 Ref

2 1.72 1.30-2.28

3 3.06 2.07-4.52

4 2.63 1.80-3.83

N stage 0 Ref

1 1.68 1.28-2.20

2+ 2.77 1.98-3.85

M stage 0 Ref

1 2.97 2.12-4.16

Grade I Ref

II 3.15 1.80-5.51

III 6.10 3.41-10.9

ER/PR Positive Ref

Negative 1.47 1.10-2.01

HER-2 Negative Ref

Equivocal 0.92 0.53-1.59

Positive 0.98 0.74-1.29

Definitive local therapy Yes Ref

No 2.07 1.49-2.88

Chemotherapy Yes Ref

No 1.17 0.86-1.56

Endocrine therapy Yes Ref

No 1.16 0.87-1.56

Delay in surgery Yes Ref

No 1.07 0.79-1.47

Delay in adjuvant therapy Yes Ref

No 0.97 0.74-1.27

Results

227

Charlson score 0 Ref

1-2 1.33 1.04-1.72

3+ 1.29 0.66-2.51

Smoking status Non-smoker Ref

Ex-smoker 1.25 0.90-1.75

Current smoker 1.10 0.84-1.45

Facility type Public Ref

Private 0.87 0.67-1.13

Deprivation Dep 1-2 Ref

Dep 3-4 1.69 1.02-2.80

Dep 5-6 1.55 0.99-2.44

Dep 7-8 1.54 0.98-2.43

Dep 9-10 1.50 0.95-2.36

Residence Urban Ref

Semi-urban 0.87 0.68-1.11

Rural 0.83 0.63-1.10

Hazard ratios for breast cancer-specific mortality with sequential adjustments for factors of

interest are provided in Table 41. The baseline model is adjusted for age and year of diagnosis,

whereas the fully adjusted model adjusts for mode of diagnosis, tumour characteristics,

treatment including delays, comorbidities, and demographics. Results are presented for all

stages of disease and then stratified by stage I/II and III/IV.

All stages

When all invasive cancers are considered, Māori women had a significantly higher hazard of

death, which was more than double that for NZ European women at baseline (HR=2.02; 95%

95% CI, 1.59–2.58); but was explained almost fully, and was no longer significant (HR=1.07;

95% CI, 0.80-1.44) after full adjustment.


228

Table 41: Hazard ratios for breast cancer-specific mortality risk in Māori compared with NZ European women with stepwise adjustment for demographics, screening

status, disease factors, treatment factors, patient factors and healthcare access

Characteristics Overall

HR ( 95% CI)

Stage I & II

HR ( 95% CI)

Stage III & IV

HR ( 95% CI)

Unadjusted 1.81 (1.43-2.30) 0.99 (0.62-1.58) 1.96 (1.47-2.61)

Model A (Baseline - adjusted for age and year of diagnosis) 2.02 (1.59-2.58) 1.20 (0.75-1.93) 2.09 (1.56-2.80)

Model B (Model A + Screening status) 1.86 (1.46-2.37) 1.14 (0.71-1.93) 2.00 (1.49-2.67)

Model C (Model B + Cancer stage at diagnosis [TNM]) 1.48 (1.15-1.93) 1.08 (0.67-1.74) 1.67 (1.23-2.27)

Model D (Model C + Cancer biological factors)

ER/PR, Grade & HER-2 1.40 (1.09-1.81) 1.15 (0.71-1.88) 1.49 (1.09-2.05)

Model E (Model D + Treatment)

Completion of definitive local therapy 1.31 (1.01-1.69) 1.18 (0.72-1.92) 1.40 (1.02-1.93)

Use of systemic therapy a 1.26 (0.97-1.64) 1.12 (0.69-1.83) 1.35 (0.98-1.87)

Delay in surgery or adjuvant therapy b 1.25 (0.96-1.63) 1.13 (0.69-1.85) 1.41 (1.01-1.95)

Model F (Model E + Patient factors)

Comorbidity index score 1.20 (0.92-1.57) 1.08 (0.66-1.78) 1.33 (0.95-1.86)

Smoking 1.16 (0.88-1.53) 1.10 (0.66-1.85) 1.24 (0.88-1.76)

BMI 1.11 (0.83-1.48) 1.12 (0.66-1.90) 1.16 (0.81-1.66)

Model G (Model F + Healthcare access factors

Socioeconomic deprivation 1.10 (0.83-1.47) 1.07 (0.63-1.84) 1.16 (0.81-1.67)

Urban / rural residency 1.09 (0.82-1.46) 1.08 (0.63-1.84) 1.16 (0.80-1.67)

Public / private treatment 1.07 (0.80-1.44) 1.12 (0.65-1.93) 1.11 (0.77-1.61)

a chemotherapy and endocrine therapy, b delay in surgery or delay in initiating chemotherapy or radiation therapy

Discussion and Conclusions

229

Differences in mode of diagnosis (screen vs. non-screen) contributed approximately 15% of

the survival disparity. Cancer stage at diagnosis accounted for approximately a third of the

survival disparity between Māori and NZ European women, with adjustments for tumour,

lymph node and metastatic status reducing the hazard ratio from 1.86 to 1.48 (Table 40). The

contribution from cancer biological characteristics was relatively small at approximately seven

percent. Differences in treatment including use of definitive local therapy, systemic therapy,

and delays in treatment contributed approximately 15% to the total survival disparity, with a

reduction in hazard ratio from 1.40 to 1.25. Patient characteristics including medical

comorbidities, smoking and BMI contributed a further 15% to the survival disparity. In this

model, contributions from healthcare access factors were minimal (approximately 2-3%) after

adjusting for patient, tumour and treatment factors. Factors included in this final model (Table

41) together accounted for approximately 95% of the observed survival disparity between

Māori and NZ European women, and Māori women were only 7% more likely to die from

their breast cancer in the fully adjusted model.

A second model was performed with a different sequence of variable introduction (Appendix

10). For this model, healthcare access characteristics were introduced first to the baseline

model, followed by rest in the same sequence. In this model, healthcare access factors

contributed approximately 20% to the survival disparity (reduction in hazard ratio from 2.02 to

1.80). Screen-detection explained approximately 20% (HR reduction from 1.80 to 1.60) and

stage at diagnosis 25% (HR reduction from 1.60 to 1.35) of the survival disparity. No

substantial differences in hazard ratio reductions were observed for patient factors (15% vs.

16%), tumour biology (7% vs. 8%) or treatment (15% vs. 14%) between the two models.

Stage specific

In the fully adjusted model for stage I/II, Māori women had a higher hazard ratio for breast

cancer mortality (HR=1.20), which however, was substantially lower than the hazard ratio for

all cancers. Adjustments for patient, tumour, treatment and healthcare access factors resulted

only in a reduction of approximately 40% (HR reduction from 1.20 to 1.12), which was

proportionately a much smaller reduction than for all cancers.

Compared with early stage cancers, the survival disparity for Māori compared with NZ

European women for advanced staged cancers (stage III/IV) was approximately five-times

higher, at a baseline hazard ratio of 2.09. Similar to the model for all stages, adjustments for


230

patient, tumour, treatment and access to healthcare factors resulted in approximately a 90%

reduction in survival disparity with a hazard ratio decline from 2.09 to 1.11. Compared with

the overall model, greater contributions were observed for tumour biology and patient factors

(approximately 15% each), while smaller contributions were seen for screening and treatment

(approximately 8% and 1% respectively).

Discussion:

In this population based cohort study of New Zealand women with breast cancer, Māori

women were observed to have a significantly poor cancer specific survival rate, and an age

adjusted risk of death from breast cancer, which was more than double that for NZ European

women. More advanced stage at diagnosis appeared to be the major factor contributing to

excess breast cancer mortality in Māori, while other factors including comorbidities, smoking,

obesity and differences in treatment made significant contributions. Differences in cancer

biological characteristics contributed minimally to the survival disparity overall and for early

stage cancer, while for advanced cancer a substantial contribution was observed. Together

these factors explained almost all the observed breast cancer survival disparity between Māori

and NZ European women.

Most causes of inequity in breast cancer survival between Māori and NZ European women are

due to health service inequities including differential access to screening, and access,

timeliness and quality of breast cancer treatment. Access to screening is the likely main

determinant of inequities in stage at diagnosis between Māori and NZ European women which

was responsible for about 40% of the survival inequity. Despite the gradual improvement in

mammographic breast cancer screening coverage over the last decade, coverage for eligible

Māori women continues to be significantly lower than that of NZ European women, and the

target coverage of 70% (39). In addition, women with screen detected cancers in this study

appeared to have a significantly better survival even after adjusting for stage and tumour

biological characteristics. Higher likelihood of cancers with a relatively favourable prognosis

among screen detected women might explain some of it, but other confounders including less

variation in access, timeliness and quality of breast cancer care for screen versus symptomatic

pathways of care might also have contributed.

A greater survival disparity was observed between Māori and NZ European women with

advanced staged cancer (HR=2.09) than for early staged cancer (HR=1.20). It seems that


231

advanced stage at diagnosis for Māori women was associated with poorer survival not only

through the advanced nature of the disease itself, but also through other associated adverse

healthcare, patient and tumour characteristics. Greater levels of comorbidity, high BMI and

smoking for Māori versus NZ European women accounted for a quarter of the survival

disparity between Māori and NZ European women with advanced disease, while for early

stage disease the contribution was less than 15%. The greater impact of comorbidity on

reduced survival in Māori than in NZ European women for more advanced cancer indicates

that Māori women with advanced cancer may be more likely to receive sub-optimal treatment.

This theory is further supported by a previous study on colon cancer where comorbidity was

reported to be a major risk factor for sub-optimal therapy in Māori compared with non-Māori

patients (156).

Inequities comorbidities, obesity and smoking result from differential access to determinants

of health such as income, education and housing as well as access to timely, high quality

health care (376). While health care services can only play a small part in eliminating unfair

social determinants of health, they can mitigate the effects of poverty as a determinant of

access to transport to health care, timely access to health professionals, access to

pharmaceuticals, and access to social welfare.

Overall, differences in tumour biological characteristics appeared to be contributing minimally

to the survival disparity. However, for advanced stage cancers the contribution was

substantially greater than for early staged cancers (7% vs. 15%). This indicates that some of

the more advanced cancers diagnosed in Māori women might have been due to biologically

more aggressive nature and rapid growth of these cancers (e.g. higher grade and HER-2

positive). Ethnic differences in cancer biological characteristics and their contributions

towards ethnic differences in breast cancer outcomes have been studied previously (11, 166).

For example African American women are about 50% more likely to have more aggressive

breast cancers (86, 377), and these differences have been shown to be responsible for

approximately 15-25% of the breast cancer survival disparity between African and White

American women (83). Although biological differences in breast cancer between Māori and

NZ European women seem to be contributing to survival disparity, its impact appears to be

much smaller, and nature of biological differences seem to differ from the USA (80, 166).

Reasons for ethnic differences in breast cancer biology are largely unknown (301). It is

unlikely that there are innate genetic differences between ethnic groups resulting in more

aggressive breast cancers for some groups. Ethnic groups are not always from discrete


232

ancestry groups. For example, NZ European and Māori women descend from a wide range of

ancestry groups and each group shares common ancestry groups. It is likely that socio-

environmental factors put minority, Indigenous and other socially disadvantaged women at

risk of having more aggressive breast cancers than privileged ethnic groups (302).

Barriers to accessing healthcare and inferior quality of care received by Māori compared with

NZ Europeans have been shown to be an important mediator for poor outcomes in Māori for

many medical conditions, including cancer (67, 102, 182, 378). For instance, Māori have been

shown to be less likely than NZ European counterparts to undergo coronary revascularization

for ischaemic heart disease (183), to receive surgery for operable lung cancer (182), and were

less likely to have received a curative resection for operated colon cancer (67). Higher

prevalence of markers of poor healthcare access factors for Māori including poor

socioeconomic status, lack of health insurance and rural residency might have contributed to

these disparities (12, 235). Furthermore, there is evidence that Māori receive inferior quality

healthcare within healthcare institutions for instance with lower use of adjuvant therapy or

with longer delays for treatment (67). The New Zealand healthcare system is expected to

provide a high quality equitable care for all citizens (28). However, it seems that healthcare

structure, delivery of health services and possible institutional racism have contributed to a

relatively inferior cancer treatment being delivered to Māori compared with NZ European

patients (35, 67).

The New Zealand Cancer Control Strategy has identified decreasing inequalities in cancer

outcomes as a key objective alongside the objective of improving outcomes for the total

population (275). There are arguments for prioritizing equity over total population goals but

the policy does not reflect this sentiment in its strategies. Although the actions arising from the

New Zealand Cancer Control Strategy have seen some tangible improvements in cancer

outcomes for Māori (379), much work is still needs to be done, as breast cancer disparities

exist at multiple levels from screening to gaining access to healthcare through to treatment (39,

123). In an attempt to standardize care for all women with breast cancer, the Ministry of

Health recently published standards of service provision for women with breast cancer (59),

which is expected not only to provide standards of care but also tools to measure and audit

quality of care. Monitoring and reporting care by ethnicity against these standards will provide

valuable feedback to identify deficiencies, so remedial action through equity focussed quality

improvement can be targeted at ‘inequity hotspots’ along the breast cancer care pathway.


233

Although majority of the key variables used in this study had complete data, a few included

significant proportions of missing data. However, an analysis performed including only

complete datasets yielded results much similar to results shown, and hence unlikely to have

significantly affected the reported findings. Although regional variations in healthcare services

may have impacted on some of the observed differences, overall findings were comparable

with similar research on ethnic disparities in other cancers in New Zealand (67, 182). Hence,

our study findings are likely to be representative of the ethnic differences of breast cancer in

New Zealand.

In conclusion, this study has reconfirmed significantly worse breast cancer outcomes in

Indigenous Māori compared with NZ European women. Several healthcare access and

treatment patient, tumour, factors appeared to be contributing to this disparity. Equity focused

improvements to healthcare, including increasing mammographic screening coverage for

Māori women and providing equitable high quality and timely cancer care has the potential to

significantly improve the survival disparity between Māori and NZ European women.


234


235

Chapter 6. Discussion and Conclusions

6.1. Interpretation of results

Key findings of this study could be broadly categorized into differences observed in relation to

patient characteristics, tumour factors, treatment differences and differences in breast cancer

survival.

1. Socio-demographic and patient characteristics – Māori patients were significantly

different from NZ European patients in relation to healthcare access characteristics and

patient characteristics that included levels of comorbidity, smoking and obesity

2. Cancers – breast cancers in Māori were significantly more likely to be advance staged

at diagnosis and were more likely to carry some of the prognostic characteristics

associated with worse cancer outcomes. Further, Māori women (overall and within

screening age) were significantly less likely to have been diagnosed through screening

compared with NZ European women.

3. Treatment – Māori were significantly less likely to receive certain forms of cancer

treatments and were more likely to experience longer delays to receive cancer

treatment than NZ European patients. Once started on treatment, Māori were more

likely to be sub-optimally adherent with treatment.

4. Outcomes – Māori women were approximately 100% more likely to die from their

breast cancer (age adjusted) compared with NZ European women. Delay in diagnosis

due to lack of healthcare access and lower screening participation were responsible for

about a half of this disparity while comorbidity and treatment differences made

substantial contributions.

This section tries to bring together findings from studies mentioned under results chapter and

summarises key findings of this study. Study findings are discussed in the context of existing

literature on breast cancer survival disparities in New Zealand.


236

6.1.1 Differences between Māori and NZ European women with breast cancer

Significant socio-demographic, tumour and treatment differences were observed between

Māori and NZ European women with breast cancer included in this study. All these factors

appeared to be contributing at varying degrees towards the final survival inequity.

Differences in socio-demographic and patient characteristics

Demographics

Average age of Māori women with breast cancer was approximately six years lower compared

with NZ European women. This finding was compatible with age structures of Māori and NZ

European populations within the region and in New Zealand (Figure 32) (41).

Considering the age distribution pattern, Māori women are expected to have a lower crude

breast cancer mortality rate than NZ European women. However, the crude mortality rate was

approximately 80% higher for Māori, and age adjustment increased this disparity further, to

just over 100%.

Figure 32: Age structure for Māori and non-Māori populations in Waikato in 2006 (Source: Census

2006, Statistics, New Zealand)


237

The number of breast cancers diagnosed per year has increased for both groups in the

Waikato; number of cancers per year has approximately doubled for NZ Europeans (from 100

to 200 per year approximately) while the number has approximately tripled for Māori women

(from 15 to 50 per year approximately). This has resulted in a gradual increase in the

proportion of Māori breast cancers, which has increased from 12% to 19% compared with a

comparative reduction from 84% to 77% for NZ European women from 1999-2002 to 2010-

2012. Expansion of Māori population at a rate faster than NZ Europeans, and a gradual

increase in life expectancy in Māori which has resulted in a shift in the population age

structure of Māori would explain some of this increase in breast cancer incidence for Māori

(41).

Where relevant, all comparisons between Māori and NZ Europeans have been adjusted for

age/age category at diagnosis and year/year category of diagnosis.

Cancer stage

Stage at diagnosis was significantly more advanced in Māori compared with NZ European

women, which was consistent with previous literature (4, 9). However, the rate of metastatic

cancer in Māori was found to be more than twice that for NZ European women, which was

significantly higher than reported in previous literature (4). Under-staging of metastatic

cancer, which is more common in Māori, was the major reason for this difference between

publications based on the NZCR and the present study.

More advanced cancer at diagnosis in Māori seemed to be contributed both by a delay in

diagnosis due to delay in presentation to a healthcare facility and a due to a lower rate of

screen detected cancer compared with NZ European women. The actual proportional

contribution of these two factors (and the added provider delays) towards more advanced

cancer at diagnosis could not be analysed in this study due to the unavailability of data on

primary care consultations.

Patient factors

Rates of medical comorbidities, smoking and obesity were significantly higher in Māori

compared with NZ European women. These findings were comparable with published rates

for Māori and NZ European women (33). These patient factors were significant contributors


238

towards overall worse breast cancer survival in Māori. Poorly controlled medical

comorbidities, smoking and obesity are likely to be higher in women of lower socioeconomic

groups, in women with a low health literacy as well as in Māori (31). As there is a significant

overlap among these three categories, identifying the independent effect of ethnicity is

complex. Irrespective of the origin, the root cause for these disparities lie in the differential

access to determinants of health and differential access to quality healthcare.

Differences in tumour characteristics

Tumour biology was an area where there were significant myths and controversies. Lack of

proper local data, and hence trying to extrapolate based on findings from countries including

the USA has created a myth among some physicians that Māori have an inherently worse

tumour biology compared with NZ European women. A study based on the Auckland Breast

Cancer Register by Weston and colleagues have supported this theory, while others including

McKenzie et al and Dachs et al have refuted such claims (70, 74, 169). Our observation was

that Māori do have significantly higher rates of certain biological characteristics (e.g. HER-2

positive cancers), but not others (e.g. triple negative cancer) associated with worse outcomes.

Regardless, the overall contribution of these differences to the mortality disparity was

minimal, in the region of approximately 5%.

Differences in treatment – provision and adherence

Assessment of quality of treatment is complex. Some of the areas of quality are highly

subjective. Some include several steps during each treatment pathway which are difficult to

capture and hence are not documented. Only a few selected key areas of treatment were used

for analysis of treatment disparities and their contribution towards the survival disparity. These

parameters included timeliness and use of and adherence with (endocrine therapy only)

treatment. Other areas of treatment including referral for adjuvant therapy, oncology decision

process, and adherence with and completion of radiotherapy and chemotherapy were not

analysed. It is likely that similar disparities do exist in those areas which might have

contributed to the survival disparity. Although all these characteristics were not included, the

key parameters used in this study are expected to have captured most of the key treatment

differences contributing to the survival inequity.


239

Overall, Māori women appeared to have received an inferior quality of breast cancer care

compared with NZ European women. Longer delays were observed for receipt of primary

surgery as well as for adjuvant therapy. Māori were significantly more likely to have

undergone mastectomy for cancers that were suitable for breast conservation, but the rate of

post-mastectomy reconstruction was only a third that of NZ European women. Use of

radiotherapy and endocrine therapy were lower for eligible Māori compared with NZ

European women, although no such difference was observed for chemotherapy. Once started,

adherence to treatment was significantly lower in Māori than NZ European women, as

observed with lower adherence to adjuvant endocrine therapy. Although patient choice seems

to have been a contributory factor for some of these differences (e.g. higher mastectomy rate

in Māori), most of these inequities appeared to have been driven by healthcare access,

including lack of access to private healthcare for Māori, and healthcare delivery process.

Outcome differences

The risk of mortality from breast cancer in Māori was twice that for NZ European women

after adjusting for age and year of diagnosis. Survival disparity was significantly smaller for

early staged cancer (20%) compared with advanced staged cancer (109%).

Lower screening coverage and more advanced stage at diagnosis

As discussed previously, Māori women had significantly more advanced breast cancer at

diagnosis compared with NZ European women. This disparity contributed to approximately

40% of the overall survival disparity between Māori and NZ European women. Lower

proportion of screen detected cancer contributed an additional 15% to the survival disparity

which appeared to be independent of the stage at diagnosis. Inclusion of a higher proportion of

cancers with better prognosis including over diagnosis of clinically indolent low grade cancers

with screening mammography might have been responsible for some of this difference. Access

to healthcare and levels of health literacy are likely to have been better for women who

participated in screening and these might also have contributed to relatively better prognosis

for women with screen detected cancer. This is further supported by the lack of a survival

disparity observed for screen versus non-screen cancer between Māori and NZ European

women. Perhaps more importantly, this shows that Māori women with breast cancer do as well


240

as, if not better than NZ European women provided that they are diagnosed early and provided

with equitable quality of care.

Stage at diagnosis contributed by a third towards the survival disparity seen for both early

staged (stage I/II) and advanced staged cancer (stage III/IV) between Māori and NZ European

women. Differences in rates of screen detected cancer contributed another third towards

survival disparity for early staged, but was less than 10% for advanced cancers. However, only

a small proportion of advance staged cancers were included under screen detected category.

Comorbidity, obesity and smoking

Patient factors contributed approximately 15% towards the survival disparity. This

contribution was much greater for advanced stage cancers compared with early staged cancers

(25% versus <5%). Other adverse characteristics including lower health literacy and lower

socioeconomic state which are more common among these patients might have contributed for

this greater contribution seen for women with advanced staged cancer.

Quality of healthcare –provision, delays and adherence

Treatment disparities appeared to be contributing to approximately 15% of the survival

disparity between Māori and NZ European women. As only a few key areas of treatment

disparities were analysed, we may have underestimated the actual contribution. However, the

underestimation, if there is one, is likely to be minimal as key parameters used in the analysis

probably would have captured effects of some of the unmeasured parameters (e.g., use of

chemotherapy and radiotherapy are likely to have captured differences in rates of referral,

though we have not analysed differences in referrals for oncology treatment).


241

6.2. Why do these disparities exist?

Results from this study has shown that almost all, if not all, breast cancer survival inequity

between Māori and NZ European women is due to modifiable risk factors, of which the most

important was differences in timely access to quality healthcare. These findings are largely

comparable with previously reported survival disparities for bowel and lung cancers in New

Zealand (67, 182). Cancer survival disparities observed between Indigenous/ethnic minority

and other populations are not unique to New Zealand. Many other countries including

Australia and the USA experience similar or sometimes worse cancer survival disparities (80,

81, 380).

Why do Māori have less access to timely and quality healthcare? Finding an answer to this

question is complex as these disparities are brought about by a series of patient and healthcare

service characteristics, as well as due to certain incompatibilities between the two. It is further

complicated by other covariates including lower socioeconomic status, geographical location

and health literacy which also contribute to health inequities.

As discussed in the literature review these factors could be categorized broadly into healthcare

system, healthcare process and patient characteristics for ease of understanding, despite the

significant overlaps observed among these categories.

6.2.1 Provider and healthcare system characteristics

Structural aspects of healthcare delivery appeared to be the main driver of Māori – NZ

European healthcare inequities. While some of these inequities are due to direct effects,

additional effects are mediated through socioeconomic and geographical differences between

Māori and NZ European women.

Findings from this study have highlighted some of the areas where the healthcare system has

failed to provide an acceptable and equitable level of cancer care for Māori compared with NZ

European patients. Māori were observed to experience inferior quality of care at almost each

step of the cancer care pathway, and this indicates the widespread nature of disparities

throughout the health service. As Māori tended to be diagnosed with more advance staged

disease compared with NZ European women, inferior quality of care including longer delays

likely have had a significant impact on the survival inequity.


242

Population groups who bear the burden of health services inequity enter the health system with

additional disadvantage of inequities in determinants of health such as income, education,

healthy housing, healthy employment and living in safe communities. For non-dominant or

Indigenous ethnic groups, inequities in access to social determinants of health and to quality

health care are compounded by a much greater exposure to racism. This is a major risk factor

for poor health as it impacts on access, timeliness and quality of health care.

The role of health services in creating and maintaining health inequities is well known (381).

An increase in inequities while increasing ‘health for all’ is a well-documented unintended

consequence of interventions aimed at improving population health (382). A strategic

approach that is equity focussed is imperative if the dual goals of achieving equity and

improving health for all are to be realised. A focus on improving total population health can

result in improved health overall while at the same time increasing inequities. This is because

total population health can improve even when there is a large growth in health inequities

between a majority and a minority population. The adoption of evidence based approaches for

achieving equity in cancer mortality by New Zealand health services has been inconsistent, not

well monitored and not prioritised over improving total population health. As a result, major

inequities continue to exist for a spectrum of medical conditions both by ethnicity and


As with other national systems and institutions, the New Zealand healthcare system also is

designed by and follows the European pattern. Although recent changes to the healthcare

structure and system have incorporated some Māori governance and values, these remain

relatively minor. Cancer specialist services are exclusively available through mainstream

health services dominated by a European culture, by providers with European values and

beliefs. Such a system might neither be completely acceptable nor comfortable for Māori

patients, and might have created a barrier for Māori to access optimum level of cancer care.

Cultural differences between healthcare system and patients, interfering with optimum cancer

treatment has to be recognized as different from culture, values or belief of a patient group

restricting optimum cancer care (96, 191). For instance, certain beliefs among some

ethnic/cultural groups may prevent these patients from accessing readily available healthcare

services. All ethnic groups carry a cultural identity and will be at ease receiving healthcare

from a culturally compatible healthcare system, similar to what NZ Europeans experience in

New Zealand healthcare system which largely follows a European pattern. On the other hand,


243

the healthcare system that is designed and functions along values of Europeans may create an

environment that differs from expectations of Māori patients.

Many local and international researches have reported on the influence of cultural and/or

ethnic identity of patients towards physician decision making processes (96, 191). These may

be driven on at times by discrimination, or perhaps more commonly due to a lack of

understanding of values of a particular culture or ethnicity. In the context of public healthcare

system, there is an unequal distribution of power between clinicians and patients, in favour of

clinicians. This may influence physicians to provide an inferior quality of care for patients

who belong to minority ethnic or cultural groups. In the fee levying private health system,

such disparities are less likely to exist as the physician patient relationship is driven primarily

through remuneration. However, only a small proportion of Māori are able to afford private

sector care and this has made them more vulnerable as they have to depend almost entirely on

public healthcare system for healthcare needs.

This study did not include details of physician decision making processes or patient-physician

interactions. Hence, possible differences of these characteristic between Māori and NZ

Europeans were extrapolated as possible reasons for observed disparities in cancer care,

supported through existing local and international literature (96, 191). For instance, Hill and

colleagues have reported that for stage III colon cancer, Māori and non-Māori patients are

referred at equal rates for chemotherapy (181). Yet, chances of chemotherapy being offered

for Māori for these cancers were significantly lower than for non-Māori. This suggests that

physicians are less inclined to recommend chemotherapy for Māori, possibly due to a belief

that Māori are less likely to benefit from chemotherapy or due to the belief that Māori are less

suitable for chemotherapy. It is unclear whether such discrepancies were existent for breast

cancer in the Waikato as we did not observe significant differences in rates of chemotherapy

between Māori and NZ European women. However, we did not have adequate data to

ascertain possible ethnic differences in rates of referral or being offered with chemotherapy.

6.2.2 Patient factors including socioeconomic factors

This study observed that rates of medical comorbidities, smoking and obesity were

significantly higher in Māori compared with NZ European women with breast cancer. All

these are factors well-known to be associated with an increase in breast cancer mortality (146,

377, 383). As expected, these factors were found to be contributing to approximately 15% of


244

the survival inequity. These patient related factors appeared to be contributing to higher breast

cancer mortality mostly through interfering with optimum cancer treatment.

Higher levels of comorbidity, smoking and obesity, as mentioned previously, arise as a result

of poor health literacy and differential access to determinants of healthcare including primary

and preventative healthcare services. These very same markers also contribute to delays in

diagnosis, lower rates of acceptance and adherence with treatment, which further increase

likelihoods of poorer breast cancer outcomes for these women.

There were some instances of patient choice influencing differences in treatment between

Māori and NZ Europeans. For instance Māori were more likely to opt for a mastectomy for a

cancer suitable for breast conservation. However, this is a scenario of two treatment options of

equal efficacy, and this difference is likely to have been driven by other factors including

difficulties in accessing radiotherapy after breast conservation. Equal rates of definitive local

therapy observed for early breast cancer between Māori and NZ European women further

support this. We did not have information to identify the impact of patient choice on use of

other treatment including chemotherapy and radiotherapy. It is unclear whether patient choice

has been a reason, especially in light of previous conflicting reports published based on lung

and bowel cancer. Patients declining treatment was a major factor for differences in surgical

treatment for lung cancer while patient choice had no impact on treatment differences

observed for colon cancer (67, 182).

Both Māori ethnicity and lower socioeconomic state appeared to be associated with poor

access to healthcare, an inferior quality of treatment and ultimately, with poor breast cancer

outcomes. Treatment and outcome disparities between Māori – NZ European groups tended to

be significantly greater than disparities observed between affluent and deprived

socioeconomic groups. This pattern of breast cancer treatment and outcome disparities

associated with ethnicity and socioeconomic status are compatible with treatment and outcome

disparities for breast and many other cancers reported from New Zealand and internationally

(4, 82). Underlying reasons associated with these inequities appear to be barriers to access

healthcare and inferior quality of care experienced by Māori and women of lower

socioeconomic groups. Lower socioeconomic status in Māori patients was a strong driver for

lower access and poor quality healthcare, but explained only part of the inequity. This is

further supported by the disparities observed within the same socioeconomic category between

Māori and NZ European women.


245

6.3. How can we provide Māori women with better and equitable

healthcare?

As the extent of breast cancer outcome inequity between Māori and NZ European women, and

some of the key drivers behind inequities have been ascertained, the following section

attempts to discuss ways to correct such disparities.

6.3.1 Identifying the problem and its underlying reasons

Breast cancer survival inequity between Māori and NZ European women was well known for

over three decades and some of the underlying causes including delay in diagnosis were also

known during this time (4). Current study has added to this knowledge base, first by

confirming the validity of previous findings and second, by shedding light on previously

unknown disparities, including disparities in breast cancer treatment. Many of the objective

differences in cancer care and cancer outcomes identified in this study, as well as in many

previous studies, were due to cumulative effects of many subjective and objective factors

acting throughout the cancer care pathway. Some of these factors are difficult or impossible to

measure. Hence, many of the proposed changes are based not only on findings from the

present and previous New Zealand studies, but also from personal experiences of clinicians

and research findings from other countries.

6.3.2 Providing equitable and acceptable care

Addressing the issue of providing an equitable quality of healthcare for disadvantaged

populations including Māori and low socioeconomic groups requires changes in several

aspects of the healthcare system. First, it requires a strong focus on providing equitable care.

Second, those strategies that have been shown to be effective locally as well as internationally

need to be recognized, and implemented in a manner that suits the local context. Finally, the

performance of implemented strategies needs to be evaluated and monitored by ethnicity to

identify effectiveness and deficiencies.

In this regard, Mackenbach has described four possible targets to intervene in order to reduce

ethnic and/or socioeconomic inequalities in health (384). In this framework (Figure 33) health

inequalities may be reduced by targeting:


246

1. Underlying social and economic determinants of health

2. Factors that are intermediate between socioeconomic determinants and health, such as

behaviour, environment and material resources

3. Health and disability support services

4. The feedback effect of ill health on socioeconomic position.

Figure 33: Four possible targets for interventions to reduce ethnic/socioeconomic inequalities in health

(Adapted from Mackenbach et al.)

Based on the Mackenbach model, the Ministry of Health has developed an intervention

framework model (Figure 34) which provides a guide for the development and implementation

of comprehensive strategies to improve health and reduce health inequalities. Step three of this

framework has direct relevance to interventions that are proposed based on the current study.

In broad terms, these include improving access to care, improving care pathways and targeting

to improve health through a population approach. Further, the New Zealand Cancer Control

Strategy has also identified reducing ethnic inequalities in cancer outcomes as one of its two

main objectives (275). Together, these provide the necessary policy support at governmental

and ministerial level to reduce ethnic inequalities in cancer outcomes. Within this context, a

strong focus on equity should be established and maintained at all levels of cancer including

cancer prevention and treatment which may help reduce cancer inequalities between Māori

and NZ Europeans.

1. Ethnic/socioeconomic status

2. Intermediary factors

3. Health problems

4. Feedback effect


247

Figure 34: Intervention framework to improve health and reduce inequalities (Source: Ministry of

Health)


248

Increasing the participation of Māori in healthcare provision is one of the recognised effective

strategies to reduce ethnic inequalities. Māori involvement should be targeted at all levels of

cancer care including cancer prevention, control and care of established cancer. Not only that,

participation of Māori in policy decision processes will ensure that healthcare policies are

acceptable for Māori, which would result in greater levels of acceptance and access to

healthcare in Māori. However, under-availability of suitable and qualified Māori personnel has

been a significant restriction hindering attempts at improving Māori participation within the

healthcare sector. Providing more opportunities for Māori for healthcare sector related training

and employments should be targeted to achieve greater Māori participation.

A regular evaluation of different policies, programmes and interventions will help identify the

performance of these strategies and where changes are required to achieve better

performances. Available frameworks such as Health Equity Assessment Tool (HEAT) could

help anticipate and address likely gaps in services and programmes (385) while guidance tools

such as the Medical Research Council of the UK guidance on Evaluating Complex

Interventions would help measure performances against objective tools (386). Evaluation of

performance against objective criteria allows for comparison of performance of different

strategies and same strategy within different contexts and healthcare institutions. This

potentially would allow deficiencies to be readily recognized and to rectify deficiencies

through a feedback process. Finally, this will also provide accountability for the system and

the providers to achieve equity in healthcare services.

Health system

Health service funding policies have direct impacts on health inequalities. For instance, this

study has shown that the proportion of Māori receiving cancer care from the private sector was

only a quarter that of NZ Europeans. Hence any changes to health policy that reduces public

sector healthcare funding would likely further exacerbate health inequalities between Māori

and NZ European women. Healthcare expenditure has been increasing exponentially,

especially over the last decade, and it is unrealistic to place the complete responsibility on the

government to provide best quality cancer care for all citizens. On the other hand, a better

developed private healthcare sector could help to reduce existing workload and congestion in

the public sector which could help public sector to provide better quality and timely cancer

care, within existing funding caps. Although a balance between the two may be difficult to


249

achieve, it could potentially create a better environment for delivery of optimum of cancer care

for a majority, if not all cancer patients.

Changes to the workforce development should target greater Māori provider participation and

providing training for all healthcare providers to create awareness on issues relating to equity

of care and cultural safety. Such changes will improve the understanding of healthcare

providers on needs and expectations of Māori patients who seek care from mainstream

healthcare services.

As discussed several times the importance of increasing breast cancer screening coverage for

Māori women cannot be over-emphasized. Several factors including higher background

incidence and higher rates of obesity means that Māori women stand a greater chance of

benefitting from a greater screening coverage, more than NZ European women (40). To this

end, commendable efforts and the dedication of the BSA programme has resulted in an

increase in Māori coverage by over 50% during the last decade. However, the existing ethnic

and regional variations in screening rates highlight the need for further efforts, and perhaps use

of novel strategies to increase screening coverage. These changes will be especially beneficial

to regions such as the Midland region that includes the Waikato, which consistently has

reported the lowest screening coverage in Māori (39). Increasing screening coverage in the

Midland region poses greater difficulties due to remoteness of many regions, general lower

socioeconomic status of its population and higher than average Māori population (41). Greater

area coverage and increasing the frequency of mammographic screening provided through

mobile screening units are some of the strategies which have shown to be effective, and hence

have the potential to increase rates of screening. Further, greater Māori provider participation

in the BSA programme should be targeted which has shown to improve communication and

make screening participation more acceptable for Māori women (143). Significant quality

variations in process of invitation for screening are another issue that need addressing (387).

Improving and standardizing the process of invitation appears to be an area with a greater

potential to improve rates of screening for Māori. For instance a personal letter or a personal

phone call or a personal invitation through the general practitioner has been shown to

significantly increase the screening participation compared with the common invitation letter

which is presently used by the BSA programme (388).

Improving the geographical accessibility of cancer care services is another issue that was

observed to have significant effects not only for Māori, but also for all rural dwelling women.

Recent changes to the health policy and cost cutting measures have seen many cancer services


250

being confined to centralized tertiary care units (37). Although it is logistically impossible to

provide all specialist cancer care through regional units, increasing the distribution and

frequency of oncology outreach clinics could improve access and acceptance of cancer therapy

for many rural women. Implementation of such a strategy in the Waikato area is likely have a

greater impact where oncology services are focused at a central unit that covers needs of

cancer patients spread over an area of over 20,000 square kilometres.

Although many of the key factors contributing to ethnic disparity in breast cancer mortality

has been documented several key areas remain unexplored. These include understanding the

mechanisms of observed disparities and identifying the efficacy and cost effectiveness of

different strategies to correct identified disparities in cancer care.

Healthcare processes

Structural organizations of healthcare services create the necessary background while

healthcare process further promotes ethnic inequities in cancer. Women with breast cancer are

required to go through a complex diagnostic and treatment processes which may extend well

over a few years. Complexities of these processes combined together with social, economic

and psychological impacts of cancer could result in many women being unable to complete

these treatment processes in a manner that enable them to achieve optimum outcomes (389). A

better healthcare delivery process could minimize complexities of care and should provide

support for women with cancer for them to deal with social, economic and psychological

issues through the treatment process.

Coordinated care and patient navigation are some of more recently introduced strategies that

have shown to help women navigate through the complex cancer care pathway (229, 390).

Although we could not directly show the impact of Cancer Care Coordinators (CCC) in this

study, it is likely that a major portion of the reduction in treatment disparities observed over

last 3-year period of the study is attributed to their work. Still, care provided through two

CCC’s appeared to be inadequate, as any given point of time, over 100 women with breast

cancer are undergoing active treatment in the Waikato. The government has recognized the

importance of care provided by the CCC’s and has provided funding to all DHB’s to recruit

CCC’s to manage common cancers in New Zealand (229). Although these efforts should be

commended, there is a greater need for CCC’s especially for breast cancer as the numbers are

greater and treatment is likely to be more complicated and protracted.


251

Providing support for women to deal with psycho-social issues during the treatment processes

will be of greater benefit for Māori as they are more likely to be socioeconomically deprived,

hence more likely to encounter greater social and financial barriers. Further, many Māori

women expect to be with their whānau during the process of treatment. As such, provision of

transport and temporary accommodation may have to include whānau members. At present

some of these services are available, but many Māori women seem to be unaware of the

availability. Hence, there should be an active approach to inform women of the availability of

these services and process of obtaining such services. CCC’s who are expected to act as the

single point of contact for these women has the potential to provide such services, but as

mentioned above, means an additional workload and may require an increase in staff.

As discussed previously, provision of training for all healthcare staff on values and

expectations of Māori patients and maintaining cultural safety will develop a better

relationship between Māori patients and healthcare providers. Dedicated Māori social workers

could supplement better communication through trust and better relationship building between

healthcare services and Māori patients. This will have a significant impact not only on routine

cancer treatment processes, but also on palliative care where there has been a greater

reluctance in Māori to accept care in a mainstream setup away from their whānau (391).

Patient factors

Breast cancer outcome inequalities are driven primarily by health system and healthcare

processes which are either not provided, or provided in a manner that is not acceptable or

comfortable to a patient group. Lack of access, especially for determinants of healthcare would

manifest as higher rates of smoking, obesity and comorbidities as was observed among Māori

in this study. Equal access for healthcare services could reduce rates of poorly controlled

comorbidity, smoking and obesity in these women. Further, impact of many of these

conditions on breast cancer mortality could be minimized by better control of comorbidities or

by giving up smoking after the diagnosis of cancer (392). However, these interventions require

time and effort from the healthcare providers which at times could be quite daunting.

On a broader context, minimizing the impact of these factors on outcomes from breast and

other cancers require a better access for determinants of health for Māori women. This

includes better education and health literacy, through which many healthcare improvements

could be achieved for the whole Māori population.


252

6.4. Strengths and Limitations – How valid are the findings?

The validity of a research is dependent upon its design (appropriate to answer the study

question), its conduct and analysis. Next section looks at key strengths and key limitations of

this study and discusses the possible errors that have been induced through bias, confounding

and chance.

In summary, the key strengths of this study include the completeness of the study sample,

comprehensiveness of the data and relatively long follow up. Limited sample size that

restricted study power of some the analysis, impact of unmeasured/under-measured

confounders and missing data were identified as key limitations.

6.4.1 Study design

Completeness of the population based sample of women included in this study, richness of the

data which were collected prospectively (i.e. the WBCR data) or through a manual review of

patient clinical records (i.e. retrospectively collected data) and the high proportion of Māori

and rural women included (compared with rest of New Zealand) are the primary strengths of

this study. This allowed for comprehensive comparisons of socio-demographic tumour and

treatment characteristics by ethnicity, and to analyse for impacts of these differences on breast

cancer outcome inequities between Māori and NZ European women. A further strength is the

relatively long follow up of women included in this study, which has provided data for a

robust survival analysis. Proportional hazard modelling provided the opportunity of

identifying the impact of each factor of interest on final outcome disparity between Māori and

NZ European women.

The major limitation of this study was the limited sample size. Although this study included

data for all women with breast cancer diagnosed in the Waikato over a period of 14 years, the

Māori sample size was relatively small with only 429 invasive cancers. This number is small

especially compared with previous studies carried out with routinely collected administrative

data which included many thousands of patients (4, 9, 13, 70). Although the inclusion of

women diagnosed over a 14-year period provided a relatively larger sample size and a longer

follow up, it has also introduced a degree of heterogeneity. This heterogeneity appeared to be

greatest for breast cancer treatment, where significant and major changes were observed

during this period.


253

Pacific women in NZ have also experienced an upward trend in both breast cancer incidence

as well as mortality, especially over the last two decades (7, 73). However, as Pacific women

made up only about 2% of the Waikato cohort of women with breast cancer, meaningful

analysis of association of factors and outcomes in this group was practically impossible.

Hence, all ethnic comparisons were performed between Māori and NZ European women and

Pacific women were excluded from those analysis. This was another significant limitation of

the present study.

Study data

The data included in this study were derived either from the WBCR or through a retrospective

manual clinical notes review. Both these processes included a thorough review of all clinical

information relating to breast cancer from several sources including prospectively collected

data forms, clinical records and pathology reports. Further, an audit process of completed

records in the database helped to minimize errors in data entry, and to maintain accuracy and

uniformity of data entry. As mentioned previously, this process has enabled the WBCR to

become the most complete and comprehensive breast cancer registry in New Zealand at

present (259, 260). For instance, a comparison of data included in this study versus the NZCR

identified that 12.3% of breast cancers in the NZCR to be unstaged compared with less than

1% in this study.

Review of medical records also provided the opportunity of assessing the full breast cancer

care pathway from the point of first clinical assessment through diagnosis and treatment to the

final breast cancer outcome. This information enabled us to analyse differences in the use and

timeliness of different treatments among groups of women of interest and to identify their

impacts on cancer survival.

Study power

Power of this study is restricted by the limited sample size of Māori included in the study.

Although approximately 20% of the population in the Waikato are Māori, only about 15% of

the breast cancers were in Māori due to younger age distribution in Māori compared with NZ

European women (Figure 1).


254

There were several other options for increasing the Māori sample size and increasing the study

power. These included increasing the eligibility time limit (include women diagnosed prior to

1999) or to include women with breast cancer diagnosed outside the Waikato region.

Increasing the time limit would probably have reduced the quality of data and increased the

proportion of missing data and heterogeneity. For example many women prior to 1999 did not

have hormone receptor status or tumour grade recorded in pathology reports. Further, BSA

programme was initiated in 1999 and inclusion of women diagnosed prior to 1999 would have

resulted in a significant shift in the proportions of screen versus non-screen detected breast

cancers further increasing sample heterogeneity.

Although the inclusion strategy used in the present study was not the most efficient to detect

Māori - NZ European differences due to unequal proportions of Māori and NZ European, it

has helped to minimize possible selection bias that would have included with a smaller NZ

European sample. Further, overall larger study sample allowed to identify impact of various

socio-demographic, cancer and treatment characteristics on breast cancer outcomes in the

complete cohort, and to generate some conclusions based on differences in distribution of

those characteristics between Māori and NZ European women. This was important as in some

analyses the study power was not adequate to demonstrate a statistically significant difference

induced by these characteristics on the mortality inequity between Māori and NZ European

women.

Misclassification of cause of death

Cause of death was derived based on information gathered from patient clinical record and

was supplemented with data from the National Mortality Collection. The New Zealand

Mortality Collection has a rigorous process to assign cause of death for individuals and is

known to be highly accurate, especially when compared with mortality records of many other

developed countries, including Australia and the USA (393).

For a majority of women included in this study (approximately 80%), there was sufficient

information available from clinical records to ascertain definitive cause of death. For these

patients the concordance of cause of death based on clinical records and the Mortality

collection was over 90% and, was over 95% when cause of death was categorized into breast

cancer non-breast cancer. For those 20% of women where the cause of death could not be


255

ascertained based on clinical record data (for example for a death of a woman who has not had

any healthcare service contact for several years), we had to use cause of death assigned by the

Mortality Collection. This would have included some degree of misclassification bias to

causes of death, but is likely to be relatively minor. Further we expect this bias to have

influenced both Māori and NZ European women in a non-differential manner, further reducing

the likelihood of a bias that could substantially influence final results of this study.

6.4.2 Internal validity / Bias

Misclassification of ethnicity

Misclassification of ethnicity, which is the primary exposure variable in this study is the most

significant possible misclassification. Ethnicity classification procedures included in this study

and the rationale for use of such procedures has been described previously. There were a

number of possibilities for misclassification of ethnicity as discussed below.

Inaccurate documentation of self-identified ethnicity was the first possibility. This error was

less likely for the WBCR patients as the patients themselves record ethnicity during the

consent process. However, misunderstanding of the question or mis-documentation of the

response might still have happened, though unlikely. Inaccurate documentation was more

likely for retrospective data collection as ethnicity was assigned primarily based on ethnicity

data collected by healthcare worker/s during hospital admission/s. Wide variations in wording

of the question related to ethnicity and the way in which this question was asked from the

patient have been reported to have led to significant undercounting of Māori (5). Further, on

many occasions the healthcare workers are known to assign ethnicity based on their

assumption of patients’ ethnicity without actually asking this question from the patient (394)

Second, a patient may opt not to reveal their true ethnic identity or select an ethnicity that is

different from their true ethnicity depending on the circumstance. It is known that some

patients who identify themselves as Māori within the Māori community, identify themselves

as NZ European when comes into contact with the healthcare system (395). Previous literature

on undercounting of Māori based on healthcare records has estimated the undercounting to be

as high as 30% (5, 396). However, recent improvements in ethnicity data collection has been

shown to have reduced the Māori undercounting rates significantly (179). Further, we have


256

used ‘ever-Māori’ approach for ethnicity assignment for retrospective data which would likely

have further minimized possible undercounting of Māori in this study.

Selection bias

Selection bias in this study is likely to have been minimal due to the complete nature of the

study sample. Data from seven women with post-mortem diagnosis of breast cancer and a

further 31 women with breast cancer diagnosed during 1999-2004 were unavailable, and were

not included in the study. This provided a participation rate of 98.6%. For the missing 31

women, cancer data or ethnicity was unavailable. Therefore it is impossible to ascertain

whether the proportional distribution of ethnicity or breast cancer outcomes for these women

were different from the complete study sample.

Information bias

Information biases applicable to this study include ethnicity, clinical information and

outcomes. Possible impacts of misclassification of ethnicity have already been discussed. Bias

in relation to misclassification of clinical details and outcomes are discussed next.

Misclassification of clinical details:

There is potential for misclassification of data included for patient and tumour clinical

characteristics, treatment and outcomes.

The majority of patient and clinical characteristics were collected through prospective data

collection forms and patient clinical records, which were filled by the attending physicians.

These data were further cross referred (for example with breast screening or oncology records)

where possible to improve accuracy. Still, it is likely that there might have been under-

documentation (e.g. comorbidities) or mis-documentation of some data. However, such under

or mis-documentations were less likely for treatment and pathology information as these were

invariably discussed and reconfirmed at the multi-disciplinary meetings and were documented

at multiple sites and time points during patient management.


257

Misclassification of outcomes / death

Misclassification of outcomes could have affected this study in two main ways. First, is the

fact of death (i.e. whether a particular woman is alive or deceased). Second is the cause of

death for women who are deceased.

Fact of death is likely to be misclassified for women who have migrated out of the country

after the diagnosis of breast cancer. The numbers who have migrated out of New Zealand

totalled 12 based on available records which unlikely to have influenced the results of this

study significantly irrespective of their outcomes. Misclassifying a living person as deceased is

another scenario, but the risk is expected to be negligible as details of all deceased women

were analysed individually.

As discussed above, based on our estimations, a small proportion of deceased women would

have had the cause of death misclassified. However this is likely to have impacted both Māori

and NZ European women in a non-differential manner. This likely would have had a minimal

impact on observed mortality inequity.

Ascertainment of disease recurrence to determine disease free survival is likely to have

included a greater bias. This can arise due to several reasons. First, disease recurrence is a

process rather than a discrete event as death. Hence some women might have lived with

disease recurrence without coming into contact with the health system. Some women with

recurrent disease might have migrated out of the area. Unlike death, it is practically impossible

to capture details of recurrence, especially if the woman was detected with recurrence in

private sector or in a different region. We used disease free survival as a secondary outcome

measure only for a few studies due to its higher risk of bias as discussed here and,

interpretations of such results are appropriately cautioned.

6.4.3 Confounding and estimating effects

Analyses of differences between Māori and NZ European women in relation to tumour,

treatment and outcomes were adjusted for a number of covariates. These included some direct

confounders (e.g. age) and others which were mediators of tumour characteristics or outcomes

(e.g. socioeconomic deprivation).


258

Confounders

Age and year of diagnosis were the only pure confounders influencing tumour characteristics

and outcomes in this study. All analyses were adjusted for these two variables (e.g. odds

ratios, hazard ratios) except for situations where it was necessary to use unadjusted variable

values (e.g. Kaplan-Meier survival curves). Adjusting for year of diagnosis made only minor

changes to Māori and NZ European odds and hazard ratio estimates. However age had a major

impact on these estimates due to the younger age distribution of Māori (Figure 1) compared

with NZ European women (e.g. breast cancer mortality hazard ratio increased from 1.81 to

2.02 after adjusting to age).

Other covariates / mediating variables

Other covariates or mediating variables are factors that were associated with both ethnicity

and, tumour characteristics and breast cancer outcomes. These factors represent a causal

pathway between ethnicity and outcomes of interest. Appropriate adjustments for these factors

were undertaken for the purpose of assessing their contribution to Māori and NZ European

disparities.

The actual contribution of each mediating variable might have been imperfectly or imprecisely

measured resulting in residual confounding/mediation. Misclassification or residual variation

within measured categories might have imparted a degree of residual variation.

Residual confounding was a likely factor in all scenarios where the NZDep 2006 was used as a

measure of socioeconomic status. The NZDep 2006 is an area based system of socioeconomic

deprivation, and use of this system in the current study to measure socioeconomic state would

likely have misclassified deprivation for some individuals resulting in residual confounding.

However, at a population level NZDep 2006 has been shown to be a valid proxy measure of

socioeconomic deprivation (397). Further, the associations observed in this study between

breast cancer outcomes and socioeconomic status was compatible with associations reported

previously and this further confirms the validity of NZDep as a measure of socioeconomic

deprivation.

Cancer stage at diagnosis is another instance where residual variation might have had an

influence. Almost all analyses were performed using either TNM stage groups or T and N

categories separately. However within each stage category or within each T or N stage, one


259

group might have had more advanced cancers. Although use of sub-stage categories (e.g. stage

IA, IB. etc.) was one option available to get over this issue, small numbers of Māori within

each sub-category meant that estimates of odds or hazard ratios were highly variable with

large confidence intervals making data interpretations difficult.

Unmeasured confounders / mediating variables

There were several unmeasured confounders / mediating variables in this study which are

known to influence outcomes of interest. Most of these confounders were factors related to

socio-demographics. For example no details of patient education, health literacy or insurance

status were available, and hence were not included in analyses. Patient education and health

literacy are known to impact several health related issues including recognition of symptoms,

health seeking behaviours, acceptance and adherence with treatment and follow up (11, 80).

Some of these factors were likely reasons for persisting large difference between Māori and

NZ European women after adjusting for available covariates (e.g. adherence to endocrine

therapy in Māori was 50% lower after adjusting for available covariates).

Estimating survival effects

Final survival analysis included a model to quantify impacts of different covariates on

observed Māori - NZ European mortality disparity. Covariates included in this model

explained about 95% of the observed disparity. However this model included several

limitations and hence interpretations based on this model needs caution. First, previously

described errors of misclassification and confounding were included in this model which

might have led to some degree of imprecision of estimates. Second, based on the sequence of

introduction, significant differences were observed on impacts of different covariates. For

example, when healthcare access related factors were introduced into this model before the

stage at diagnosis or treatment, the impact was approximately 10%, while the impact was

reduced to 2% when it was introduced into the model after stage at diagnosis and treatment.

Potential effects of misclassification, confounding and residual mediation are complex, and are

difficult to overcome. Adjusting for covariates, sub-group analyses, analyses only of complete

data and sensitivity analyses to a certain extent have helped to minimize or to estimate the

impact of these factors on final results. Presence of confounding, mediating variables and


260

misclassification do not invalidate the data or the results of this study; rather they point to

areas where one needs to be cautious and to interpret results within the correct context.

6.4.4 Chance and variability

This study included a limited sample size for Māori. This resulted in limited precision of

estimates with large confidence intervals which were more evident in sub-group analyses. Low

precision and lack of significance of subgroup analyses were dealt in most situations by using

multivariable logistic regression models with sub-groups included as covariates. As most of

the outcomes of interest included in this study had a large number of covariates, this resulted

in complex statistical calculations which might have led to a degree of imprecision, albeit with

tight confidence intervals. Every attempt has been made to minimize this imprecision by using

sub-group analyses where possible, by selecting only the minimum number of covariates for

multivariable models and by performing stepwise regression analyses.

Limited sample size has the risk of including a Type II error; that is, not detecting a difference

when one exists. There were several situation where differences between Māori and NZ

European women were observed but were not statistically significant (e.g. age adjusted HER-2

positivity rate was higher in Māori [20.7%] compared with NZ European [16.3%], but the

difference was statistically not significant). However, these differences are propotionately

small and are unlikely to be clinically significant.

6.4.5 External validity

This study was conducted using a regional cohort of women with the objective of identifying

differences in breast cancer characteristics and outcomes between Māori and NZ European

women in New Zealand. Therefore, the validity of study results and its generalizability to the

general New Zealand population requires a thorough evaluation.

There were some factors and outcomes which were clearly due to regional variations. For

example, longer delays in treatment (radiotherapy and chemotherapy) observed during 2003-

2005 time period was due to shortage of oncologists and other oncology staff in the Waikato

DHB. Yet, a pattern that was similar to other time periods was observed where more Māori

experiencing longer delays. Therefore, although the absolute values of these variables might


261

not be applicable to the whole country, trends and patterns observed are likely to be valid and

applicable.

Further, distribution of some socio-demographic characteristics of the study population was

different from rest of the country. The Waikato DHB includes more rural areas compared with

other DHBs and breast screening coverage rate within the Midland region (that includes the

Waikato DHB) has persistently been the lowest in the country (39). Although these factors

might have limited the generalizability of the study results, they have also provided the

opportunity of demonstrating differences in outcomes in relation to these factors. For example,

the high rural population and women residing more than 100km from the tertiary centre in the

Waikato DHB area allowed the demonstration of statistically significant lower coverage and

longer radiotherapy delay for these women.

When all above facts are considered, this study appears to be fairly representative of the New

Zealand population and likely to be valid for general New Zealand population of women with

breast cancer over the study period.

Findings from this study on breast cancer inequalities are likely to be applicable to many other

cancers and other chronic conditions within their contexts and with limitations. For example,

similar disparities in treatment and outcomes have been shown for bowel and lung cancer

(181, 182). However there were significant differences between those cancers and breast

cancer (e.g. there was no difference between stage at presentation for bowel cancer between

Māori and non-Māori). Although there seem to be general tends in cancer disparities between

Māori and NZ European, substantial and significant differences do seem to exist which are

unique to each cancer.

Applicability and relevance of this study data to inequities between Indigenous/ethnic minority

women with breast cancer in other countries should be interpreted within their contexts and

differences unique to each country. Where possible, results of this study have been compared

with similar studies from other countries. Although there were several similarities, some

significant differences were observed. For example African American women in the USA

have been documented to have breast cancers with significantly worse tumour biology which

was not observed in Māori women.


262


263

6.5. Conclusions

This study has shown that Māori women with breast cancer are almost twice as likely as NZ

European women to die from their disease. This survival inequity appeared to be due to

differential access to determinants of health (i.e., obesity, smoking, comorbidity) and cancer

care services for Māori compared with NZ European women. Such a concept is supported by

this study and many others who have shown the existence of persistent patient, physician, and

healthcare system barriers to accessing quality cancer care for Māori patients with cancer in

general, and for breast cancer (31, 34, 67, 97, 182). The central barriers that act at patient level

include low socioeconomic status, low health literacy and possible culturally mediated

attitudes that may contribute towards key final pathways that mediate poor cancer outcomes

observed in Māori women (34, 398). At provider level, gaps in training in patient-physician

communication could also constitute an unaddressed barrier to cancer care (99).

Achieving equity in breast cancer survival between Māori and NZ European women will

require a concerted effort by health care providers, with central leadership from the Minister of

Health to prioritise equity over improving the health of the total population. Reducing the

influence of non-clinical factors on the receipt of cancer treatment may provide an important

means of reducing ethnic inequities in cancer outcomes. Many interventions to improve access

to cancer services for Indigenous/ethnic minority populations have been developed targeting

one or more of the domains that drive inequities in access to quality cancer care. For instance,

educational strategies and behavioural models have been used successfully to overcome low

health literacy or attitudinal barriers, and to increase mammography screening rates (385).

Similarly, interventions targeted to physicians that rely on behavioural cues (e.g. reminders)

and education have been used to increase breast cancer screening rates (96). Māori provider

organizations and cultural safety education are some of the local examples of initiatives that

have emerged not in isolation but, rather, within a context of government policies that have

been shown to promote health status of Indigenous Māori populations (99, 100). A greater

integration of Māori providers into the healthcare service has shown not only to reduce

cultural barriers to access care but also to reduce possible institutional discrimination towards

Māori cancer patients (100).

This study has provided data on many previously unknown or lesser known areas contributing

to breast cancer survival inequity by ethnicity in New Zealand. Yet, many other areas remain

inadequately researched or completely unexplored. For example, breast cancer in Pacific

women in New Zealand has hardly been researched, although the breast cancer mortality rate


264

is known to be even worse than Māori. New data resources and improved study methodologies

are needed to better identify and quantify the full spectrum of all non-clinical factors

contributing to the mortality inequity and to develop strategies to facilitate appropriate cancer

care for all women with breast cancer. To date, there also is limited research on patient based

or physician-based interventions focussed on early diagnosis of cancer or cancer treatment,

and still fewer research based on interventions at the healthcare system or process level.

Interventions designed to enhance access also needs to be evaluated systematically to ensure

that patient outcomes are improved for all patients, but with a greater focus on disadvantaged

groups including Māori, in the most cost-efficient manner. Together, future research and

interventions hold the promise of eliminating the current ethnic inequalities in access to cancer

care and cancer outcomes in New Zealand.

Appendix

265

Appendices

Appendix 1

Multivariable logistic regression model for unstaged versus staged cancer in the New Zealand

Cancer Register


Age category (years) <40 Ref <0.001

40-59 0.99 0.50-1.99

60-79 1.06 0.53-2.15

80+ 3.32 1.59-6.97

Ethnicity NZ European Ref 0.664

Māori 1.09 0.75-1.59

Deprivation 1-2 Ref 0.704

3-4 1.08 0.65-2.16

5-6 1.07 0.64-1.81

7-8 1.04 0.64-01.71

9-10 0.86 0.51-1.45

Charlson score a 0 Ref 0.001

1-2 1.71 1.24-2.36

3+ 1.96 1.09-3.51

Therapeutic surgery Yes Ref <0.001

No 6.85 4.90-9.58

(OR – Odds ratio, CI – Confidence interval, a Charlson Comorbidity Score)


266

Appendix 2

Characteristics of women associated with stage at diagnosis and, univariate and multivariate logistic regression analyses for factors associated with early a

versus advanced b stage for screening age c women with newly diagnosed breast cancer

Total (N=1846) Early a Advanced b Unadjusted Adjusted

Characteristic n (%) n (%) n (%) OR 95% CI p OR 95% CI p

Screening status

Screen detected 1064 (57.6) 766 (72.0) 298 (28.0) Ref <0.001 Ref <0.001

Non-screen 782 (13.1) 251 (32.1) 531 (67.9) 5.41 4.42–6.63 5.30 4.32-6.49

Ethnicity

NZ European 1459 (29.3) 835 (57.2) 624 (42.8) Ref Ref

Māori 311 (79.0) 146 (46.9) 165 (53.1) 1.51 1.18–1.93 0.001 1.26 0.97-1.65 0.085

Pacific 26 (16.8) 8 (30.8) 18 (69.2) 3.01 1.30–6.97 0.010 2.32 0.94-5.73 0.068

Other 50 (1.4) 28 (56.0) 22 (44.0) 1.05 0.60–1.86 0.863 0.94 0.50-1.73 0.831

(a early stage = in-situ & stage I, b advanced stage = stages II, III & IV, c only cancers diagnosed from July 2004 onwards are included)

Appendix

267

Appendix 3

Characteristics of women associated with stage at diagnosis and, univariate and multivariate logistic regression analyses for factors associated with early a

versus advanced b stage for screening age c women with newly diagnosed invasive breast cancer

Total (N=1548) Early a Advanced b Unadjusted Adjusted d

Characteristic n (%) n (%) n (%) OR 95% CI p OR 95% CI p

Screening status

Screen detected 858 (55.4) 538 (62.7) 320 (37.3) Ref <0.001 Ref <0.001

Non-screen 690 (44.6) 181 (73.8) 59 (26.2) 4.72 3.79-5.88 4.54 3.63-5.68

Ethnicity

NZ European 1220 (78.8) 596 (48.9) 624 (51.1) Ref Ref

Māori 268 (17.3) 103 (38.4) 1.65 (61.6) 1.53 1.16-2.00 0.002 1.29 0.96-1.74 0.091

Pacific 24 (1.6) 6 (25.0) 18 (75.0) 2.86 1.13-7.26 0.027 2.24 0.84-6.01 0.101

Other 36 (2.3) 14 (38.9) 22 (61.1) 1.50 0.76-2.96 0.241 1.23 0.58-2.57 0.584

(a early stage = stage I, b advanced stage = stages II, III & IV, c only cancers diagnosed from July 2004 onwards are included, d adjusted for age category, deprivation and

residential status)


268

Appendix 4

Age and breast cancer biological characteristics at diagnosis compared between NZ European

and Māori women.

Characteristic NZ European (N=2304) Māori (N=429) p

n (crude %) Age


Age

adjusted %

ER/PR

ER+ 1884 (83.7) 84.5 334 (79.1) 80.6 0.011

ER- 366 (16.3) 15.5 88 (21.9) 19.4

ER Unknown 54 7

Appendix

269

Appendix 5

Cox regression model for factors associated with breast cancer specific mortality

Characteristic Univariate Multivariate Multivariate

(Post 2005 only)

HR 95% CI p HR 95% CI p HR 95% CI p

Ethnicity a

NZ European Ref <0.001 Ref 0.007 0.102

Māori 1.98 1.55-2.54 1.41 1.09-1.82 1.41 0.93-2.14

T stage

T1 Ref <0.001 Ref <0.001 <0.001

T2 3.33 2.56-4.33 2.21 1.67-2.92 3.91 2.11-7.23

T3 8.68 6.01-12.5 3.99 2.68-5.94 5.35 2.51-11.4

T4 18.6 13.8-25.1 4.60 3.17-6.65 5.94 2.93-12.1

N stage

N0 Ref <0.001 <0.001 <0.001

N1 2.03 1.58-2.60 1.49 1.15-1.93 1.75 1.12-2.72

N2+ 4.04 3.01-5.42 3.03 2.41-3.92 3.42 2.41-5.26

M Stage

M0 Ref <0.001 <0.001 <0.001

M1 15.4 12.2-19.4 5.04 3.09-6.71 5.32 3.55-8.01

Mode of detection

Non-screen Ref <0.001 <0.001 0.017

Screen 0.26 0.20-0.35 0.57 0.43-0.78 0.46 0.25-0.87


270

Appendix 6

Multivariable model for factors associated with delay in chemotherapy longer than 90 days a

Delay in chemotherapy >90 days

OR 95% CI p

Ethnicity

NZ European Ref

Māori 1.41 0.66-3.03 0.291

Pacific 1.10 0.22-5.58 0.908

Other 0.82 0.10-6.78 0.853

Year of diagnosis b 1.37 1.11-1.61 <0.001

Re-excision 3.96 1.93-7.70 0.001

Surgical facility type 4.89 1.78-11.3 0.001

Distance from hospital 1.16 0.90-1.49 0.245

a Adjusted for age, tumour stage, socioeconomic deprivation and comorbidity score, b Year categories

as in Table 26.

Appendix

271

Appendix 7

Multivariable logistic regression analysis for factors associated with use of adjuvant

chemotherapy for invasive breast cancer.


Māori ethnicity 1.25 0.85-1.87 0.258

Age a 0.88 0.79-0.99 0.031


ER and/or PR positive 0.43 0.28-0.65 <0.001

Deprivation 0.96 0.85-1.28 0.527

Charlson score 0.30 0.20-0.45 <0.001

Hospital type 0.81 0.58-1.13 0.217

T stage 1.24 0.99-1.56 0.058

N stage 2.01 1.64-2.48 <0.001

Grade 2.24 1.70-2.93 <0.001

LVI 1.74 1.24-2.45 0.001

HER-2 1.27 1.11-1.45 <0.001



272

Appendix 8

Characteristics associated with women undergoing mastectomy for whom details of decision

process was available versus women for whom this information was unavailable

Characteristic Total (N=751)

n (%)

Decision known

(N=600)

n (%)

Decision not known

(N=151)

n (%)

p

Ethnicity 0.535

NZ European 599 474 (79.1) 125 (20.9)

Māori 120 101 (84.2) 19 (15.8)

Age (years) 0.113

<40 40 31 (77.5) 9 (22.5)

40-49 142 115 (81.0) 27 (19.0)

50-59 180 136 (75.6) 44 (24.4)

60-69 148 118 (79.7) 30 (20.3)

70-80 155 130 (83.9) 25 (16.1)

80+ 86 70 (81.4) 16 (18.6)

Deprivation 0.392

Dep 1-2 81 71 (87.3) 10 (12.3)

Dep 3-4 73 59 (80.8) 14 (19.2)

Dep 5-6 187 150 (80.2) 37 (19.8)

Dep 7-8 217 172 (79.3) 45 (20.7)

Dep 9-10 193 148 (76.7) 45 (23.3)

Hospital type 0.549

Public 530 427 (80.6) 103 (19.4)

Private 221 173 (78.3) 48 (21.7)

T stage 0.654

T1 353 285 (80.7) 68 (19.3)

T2 398 315 (79.1) 83 (20.9)

Charlson score 0.311

0 591 466 (78.8) 125 (21.2)

≥1 160 134 (83.8) 26 (16.2)

Appendix

273

Appendix 9

Characteristics associated with women undergoing major breast reconstruction following

mastectomy for breast cancer in Waikato 1999-2012 a

Characteristic

Total

(N=1910)

n (%)

Sentinel node biopsy

(N=263)

n (%)

Adjusted

OR 95% CI p

Ethnicity

NZ European 1584 (83.0) 873 (55.2) Ref

Māori 258 (13.4) 143 (55.4) 0.81 0.57-1.14 0.237

Pacific 24 (1.3) 9 (37.5) 0.38 0.18-1.14 0.060

Other 44 (2.3) 22 (50.0) 0.53 0.25-1.11 0.094

Age (years)

<40 78 (4.1) 34 (43.6) 0.63 0.35-1.15

40-49 369 (19.3) 211 (57.2) Ref <0.001

50-59 545 (28.5) 313 (57.4) 1.18 0.85-1.64

60-69 525 (27.6) 323 (61.3) 1.04 0.74-1.45

70-80 281 (14.7) 127 (45.4) 0.59 0.40-0.90

80+ 112 (5.9) 41 (36.6) 0.38 0.22-0.68

Deprivation

Dep 1-2 200 (10.5) 123 (61.5) Ref 0.552

Dep 3-4 209 (11.0) 108 (51.7) 0.70 0.42-1.15

Dep 5-6 479 (25.1) 271 (56.7) 0.92 0.61-1.39

Dep 7-8 554 (29.0) 297 (53.7) 0.85 0.55-1.24

Dep 9-10 468 (24.5) 248 (53.0) 0.93 0.61-1.44

Hospital type

Public 1306 (68.3) 680 (52.1) Ref <0.001

Private 604 (31.7) 367 (60.8) 1.82 1.42-2.35

Year of diagnosis

1999-2002 394 (20.6) 84 (21.3) Ref <0.001

2003-2006 574 (30.0) 219 (38.2) 2.27 1.66-3.11

2007-2009 453 (23.7) 330 (73.0) 13.9 9.81-19.8

2010-2012 489 (25.6) 414 (84.7) 35.3 21.9-47.6


274

T stage

T1 1232 (64.5) 785 (63.8) Ref <0.001

T2 678 (35.5) 262 (38.6) 0.26 0.20-0.32

Charlson score

0 1626 (85.2) 918 (56.5) Ref

≥1 284 (14.2) 139 (48.9) 0.76 0.53-10.5 0.101

Appendix

275

Appendix 10

Hazard ratios for breast cancer-specific mortality risk in Māori compared with NZ European

women with stepwise adjustment for demographics, screening status, disease factors,

treatment factors, patient factors and healthcare access.

Characteristics Overall HR ( 95% CI)

Unadjusted 1.81 (1.43-2.30)

Model A (Baseline - adjusted for age and year of diagnosis) 2.02 (1.59-2.58)

Model B (Model A + Healthcare access factors

Socioeconomic deprivation 1.91 (1.49-2.46)

Urban / rural residency 1.92 (1.49-2.46)

Public / private treatment 1.80 (1.39-2.32)

Model C (Model B + Screening status) 1.60 (1.24-1.98)

Model D (Model C + Cancer stage at diagnosis [TNM]) 1.36 (1.04-1.78)

Model E (Model D + Cancer biological factors)

ER/PR, Grade & HER-2 1.31 (1.01-1.71)

Model F (Model E + Treatment)

Completion of definitive local therapy 1.24 (0.94-1.63)

Use of systemic therapy a 1.21 (0.92-1.60)

Delay in surgery or adjuvant therapy b 1.19 (0.91-1.51)

Model G (Model F + Patient factors)

Comorbidity index score 1.16 (0.88-1.48)

Smoking 1.12 (0.84-1.46)

BMI 1.07 (0.81-1.41)

a chemotherapy and endocrine therapy, b delay in surgery or delays in initiating chemotherapy or

radiation therapy


276

Appendix 11

Time trends in 5-year breast cancer survival rates by ethnicity during 1999-2012

NZ European NZ European

Māori Māori

79%

89%

73%

87%

References

277

References

1. Ministry of Health. Cancer: New Registrations and Deaths 2010. Wellington; 2013.

2. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide

burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer. 2010;127(12):2893-917.

3. Althuis MD, Dozier JM, Anderson WF, Devesa SS, Brinton LA. Global trends in breast

cancer incidence and mortality 1973-1997. Int J Epidemiol. 2005;34(2):405-12.

4. Robson B, Purdie G, Cormack D. Unequal Impact II: Māori and Non-Māori Cancer

Statistics by Deprivation and Rural-Urban Status 2002-2006. Wellington: Ministry of

Health; 2010.

5. Ajwani S, Blakely T, Robson B, Atkinson J, Kiro C. Unlocking the numerator-

denominator bias III: adjustment ratios by ethnicity for 1981-1999 mortality data. The

New Zealand Census-Mortality Study. N Z Med J. 2003;116(1175):U456.

6. Soeberg M, Blakely T, Sarfati D, Tobias M, Costilla R, Carter K, Atkinson J. Cancer

Trends: Trends in cancer survival by ethnic and socioeconomic group, New Zealand

1991-2004. Wellington: University of Otago and Ministry of Health; 2012.

7. Sarfati D, Blakely T, Shaw C, Cormack D, Atkinson J. Patterns of disparity: ethnic and

socio-economic trends in breast cancer mortality in New Zealand. Cancer Causes

Control. 2006;17(5):671-8.

8. Lethaby AE, Mason BH, Holdaway IM, Kay RG. Age and ethnicity as prognostic

factors influencing overall survival in breast cancer patients in the Auckland region.

Auckland Breast Cancer Study Group. N Z Med J. 1992;105(947):485-8.

9. Jeffreys M, Stevanovic V, Tobias M, Lewis C, Ellison-Loschmann L, Pearce N, Blakely

T. Ethnic inequalities in cancer survival in New Zealand: linkage study. Am J Public

Health. 2005;95(5):834-7.

10. Williams SM, Templeton AR, Swallen K, Cooper R, Kaufman J. Race and genomics. N

Engl J Med. 2003;348(25):2581-2.

11. Rosenberg J, Chia YL, Plevritis S. The effect of age, race, tumor size, tumor grade, and

disease stage on invasive ductal breast cancer survival in the U.S. SEER database. Breast

Cancer Res Treat. 2005;89(1):47-54.

12. McKenzie F, Ellison-Loschmann L, Jeffreys M. Investigating reasons for socioeconomic

inequalities in breast cancer survival in New Zealand. Cancer Epidemiol.

2010;34(6):702-8.

13. McKenzie F, Ellison-Loschmann L, Jeffreys M. Investigating reasons for ethnic

inequalities in breast cancer survival in New Zealand. Ethn Health. 2011;16(6):535-49.

14. Cunningham R, Shaw C, Blakely T, Atkinson J, Sarfati D. Ethnic and socioeconomic

trends in breast cancer incidence in New Zealand. BMC Cancer. 2010;10:674.


278

15. Bennett H, Marshall R, Campbell I, Lawrenson R. Women with breast cancer in

Aotearoa New Zealand: the effect of urban versus rural residence on stage at diagnosis

and survival. N Z Med J. 2007;120(1266):U2831.

16. Cunningham R, Sarfati D, Hill S, Kenwright D. An audit of colon cancer data on the

New Zealand Cancer Registry. N Z Med J. 2008;121(1279):46-56.

17. Statistics New Zealand. 2013 Census Data 2014 [cited 2014]. Available from:

http://www.stats.govt.nz/Census/2013-census.aspx

18. King M. The Penguin history of New Zealand. Auckland: Penguin Books; 2003.

19. Ministry for Culture and Heritage. Te Ara the encyclopedia of New Zealand

[Internet]Wellington [Available from: http://www.teara.govt.nz/

20. Sorrenson MPK. Māori and European since 1870: a study in adaptation and adjustment.

Auckland: Heinemann Educational Books; 1967.

21. New Zealand Cancer Control Trust. The New Zealand cancer control strategy

[Internet]Wellington: Ministry of Health and the New Zealand Cancer Control Trust;

2003 [Available from: http://www.health.govt.nz/publication/new-zealand-cancer-

control-strategy]

22. Jansen P, Smith K. Māori experiences of primary health care: Breaking down the

barriers. N Z Fam Pract. 2006;33(5):298-300.

23. Pool DI. Te iwi Māori: a New Zealand population, past, present & projected. Auckland:

Auckland University Press; 1991.

24. Orange C. An illustrated history of the Treaty of Waitangi. Wellington: Allen & Unwin;

1990.

25. Kawharu IH. Waitangi: Māori and Pakeha perspectives of the Treaty of Waitangi.

Auckland: Oxford University Press; 1989.

26. Belich J. Making peoples : a history of the New Zealanders : from Polynesian settlement

to the end of the nineteenth century. Northshore: Penguin; 2007.

27. Ratima MM. Kia uruuru mai a hauora: being healthy, being Māori: conceptualising

Māori health promotion [PhD]. Wellington: University of Otago; 2001.

28. Ministry of Health. The New Zealand health strategy. Wellington: Ministry of Health;

2000.

29. Ministry of Health. He korowai oranga: Māori health strategy. Wellington 2002.

30. Durie M. Whaiora : Māori health development. 2nd ed. Auckland: Oxford University

Press; 1998.

31. Robson. B, Harris. R. Hauora: Màori Standards of Health IV. A study of the years 2000–

2005. Wellington: Te Ròpù Rangahau Hauora a Eru Pòmare; 2007.

32. Ministry of Māori Development. Ma te Māori e puri te maimoatanga Māori: Managed

care by Māori. Wellington: Te Puni Kokiri; 1994.

http://www.stats.govt.nz/Census/2013-census.aspx

http://www.teara.govt.nz/

http://www.health.govt.nz/publication/new-zealand-cancer-control-strategy

http://www.health.govt.nz/publication/new-zealand-cancer-control-strategy

References

279

33. Ministry of Health. New Zealand Health Survey - Annual Update 2012-2013.

Wellington; 2014.

34. Ministry of Health. Tatau Kahukura: Māori Health Chart Book 2010. 2nd Edition. ed.

Wellington2010.

35. Harris R, Tobias M, Jeffreys M, Waldegrave K, Karlsen S, Nazroo J. Effects of self-

reported racial discrimination and deprivation on Māori health and inequalities in New

Zealand: cross-sectional study. Lancet. 2006;367(9527):2005-9.

36. French S, Old A, Healy J. Health care systems in transition: New Zealand 2001.

Copenhagen, Denmark: European Observatory on Health Care Systems; 2001.

37. Gauld R. Revolving doors: New Zealand's health reforms. Wellington: Institute of

Policy Studies and Health Services Research Centre, Victoria University of Wellington;

2001.

38. OECD Health Division. OECD Health Data 2011 - Frequently Requested Data. Paris,

France; 2011.

39. Page A, Morrell S, Taylor R. BreastScreen Aotearoa: Independent Monitoring Report,

Screening and assessment report of women attending BSA [Internet]. Wellington:

National Screening Unit, Ministry of Health; 2014.

40. Robson B, Stanley J, Ikeda T, Stairmand J. Independent Māori Monitoring Report 4,

Screening and Assessment July 2010 to June 2012. Wellington: Te Roopu Rangahau

Hauora a Eru Pomare, University of Otago Wellington, New Zealand; 2014.

41. Statistics New Zealand. 2006 Census Data 2007. Available from:

http://www.stats.govt.nz/

42. Ferlay J, Soerjomataram I, Ervik M, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin D,

Forman D, Bray F. GLOBOCAN 2012 v1.0, Cancer Incidence and Mortality Worldwide

Lyon, France: International Agency for Research on Cancer; 2013. Available from:

http://globocan.iarc.fr

43. Ferlay J, Shin H, Bray F, Forman D, Mathers C, Parkin D. GLOBOCAN 2008 v1.2,

Cancer Incidence and Mortality Worldwide IARC Cancer Base 2010. Lyon, France:

International Agency for Research on Cancer; 2010.

44. New Zealand Guidelines Group. Management of early breast cancer - evidence based

best practice guideline. Wellington: Ministry of Health; 2009.

45. Olivotto IA, Bajdik CD, Ravdin PM, Speers CH, Coldman AJ, Norris BD, Davis GJ,

Chia SK, Gelmon KA. Population-based validation of the prognostic model

ADJUVANT! for early breast cancer. J Clin Oncol. 2005;23(12):2716-25.

46. Wishart GC, Bajdik CD, Azzato EM, Dicks E, Greenberg DC, Rashbass J, Caldas C,

Pharoah PD. A population-based validation of the prognostic model PREDICT for early

breast cancer. Eur J Surg Oncol. 2011;37(5):411-7.

47. Cronin M, Sangli C, Liu M-L, Pho M, Dutta D, Nguyen A, Jeong J, Wu J, Langone KC,

Watson D. Analytical validation of the Oncotype DX genomic diagnostic test for

http://www.stats.govt.nz/

http://globocan.iarc.fr/


280

recurrence prognosis and therapeutic response prediction in node-negative, estrogen

receptor–positive breast cancer. Clin Chem. 2007;53(6):1084-91.

48. Wittner BS, Sgroi DC, Ryan PD, Bruinsma TJ, Glas AM, Male A, Dahiya S, Habin K,

Bernards R, Haber DA. Analysis of the MammaPrint breast cancer assay in a

predominantly postmenopausal cohort. Clin Cancer Res. 2008;14(10):2988-93.

49. Cuzick J, Sestak I, Baum M, Buzdar A, Howell A, Dowsett M, Forbes JF. Effect of

anastrozole and tamoxifen as adjuvant treatment for early-stage breast cancer: 10-year

analysis of the ATAC trial. Lancet Oncol. 2010;11(12):1135-41.

50. Rea D, Gray R, Bowden S, Handley K, Earl H, Poole C, Bates T, Dewar J, Raytor Z,

Lee M. Overall and subgroup findings of the aTTom trial: A randomised comparison of

continuing adjuvant tamoxifen to 10 years compared to stopping after 5 years in 6953

women with ER positive or ER untested early breast cancer. Eur J Cancer.

2013;49:S402-S.

51. Davies C, Pan H, Godwin J, Gray R, Arriagada R, Raina V, Abraham M, Medeiros

Alencar VH, Badran A, Bonfill X, Bradbury J, Clarke M, Collins R, Davis SR,

Delmestri A, Forbes JF, Haddad P, Hou MF, Inbar M, Khaled H, Kielanowska J, Kwan

WH, Mathew BS, Mittra I, Muller B, Nicolucci A, Peralta O, Pernas F, Petruzelka L,

Pienkowski T, Radhika R, Rajan B, Rubach MT, Tort S, Urrutia G, Valentini M, Wang

Y, Peto R. Long-term effects of continuing adjuvant tamoxifen to 10 years versus

stopping at 5 years after diagnosis of oestrogen receptor-positive breast cancer: ATLAS,

a randomised trial. Lancet. 2013;381(9869):805-16.

52. National Breast and Ovarian Cancer Centre. Clinical Practise Guidelines for the

Management of Early Breast Cancer. 2nd ed. Australia: National Health and Medical

Research Council (NHMRC), Australia; 2001.

53. Wilson R, Liston J, Cooke J, Duncan K, Given-Wilson R, Kutt E, Michell M, Patnick J,

Thompson W, Wallis M, Wilson M. Clinical guidelines for breast cancer screening

assessment. Sheffield, UK: NHS Cancer Screening Programmes; 2005.

54. National Institute for Clinical Excellence (NICE). Guidance on Cancer Services

Improving Outcomes in Breast Cancer - a manual update. United Kingdom; 2002.

55. Alexander FE, Anderson TJ, Brown HK, Forrest AP, Hepburn W, Kirkpatrick AE, Muir

BB, Prescott RJ, Smith A. 14 years of follow-up from the Edinburgh randomised trial of

breast-cancer screening. Lancet. 1999;353(9168):1903-8.

56. Chang JH, Vines E, Bertsch H, Fraker DL, Czerniecki BJ, Rosato EF, Lawton T, Conant

EF, Orel SG, Schuchter L, Fox KR, Zieber N, Glick JH, Solin LJ. The impact of a

multidisciplinary breast cancer center on recommendations for patient management: the

University of Pennsylvania experience. Cancer. 2001;91(7):1231-7.

57. Group BAIM. BreastScreen Aotearoa: monitoring report no. 1. Wellington: Hugh Adam

Cancer Epidemiology Unit, University of Otago; 2000.

58. Ministry of Health. Faster cancer treatment indicators: Data definitions and reporting for

the indicators. Wellington; 2012.

References

281

59. Ministry of Health. Standards of Service Provision for Breast Cancer Patients in New

Zealand (Provisional). Wellington 2013.

60. Khatcheressian JL, Hurley P, Bantug E, Esserman LJ, Grunfeld E, Halberg F, Hantel A,

Henry NL, Muss HB, Smith TJ. Breast cancer follow-up and management after primary

treatment: American Society of Clinical Oncology clinical practice guideline update. J

Clin Oncol. 2013;31(7):961-5.

61. Metcalfe S, Evans J, Priest G. PHARMAC funding of 9-week concurrent trastuzumab

(Herceptin) for HER2-positive early breast cancer. N Z Med J. 2007;120(1256):U2593.

62. Burstein HJ, Temin S, Anderson H, Buchholz TA, Davidson NE, Gelmon KE, Giordano

SH, Hudis CA, Rowden D, Solky AJ. Adjuvant Endocrine Therapy for Women With

Hormone Receptor–Positive Breast Cancer: American Society of Clinical Oncology

Clinical Practice Guideline Focused Update. J Clin Oncol. 2014:JCO. 2013.54. 258.

63. Bramley D, Hebert P, Jackson R, Chassin M. Indigenous disparities in disease-specific

mortality, a cross-country comparison: New Zealand, Australia, Canada, and the United

States. N Z Med J. 2004;117(1207):U1215.

64. Pearce NE, Davis PB, Smith AH, Foster FH. Social class, ethnic group, and male

mortality in New Zealand, 1974-8. J Epidemiol Community Health. 1985;39(1):9-14.

65. Pearce NE, Davis PB, Smith AH, Foster FH. Mortality and social class in New Zealand.

III: male mortality by ethnic group. N Z Med J. 1984;97(748):31-5.

66. Dockerty JD, Marshall S, Fraser J, Pearce N. Stomach cancer in New Zealand: time

trends, ethnic group differences and a cancer registry-based case-control study. Int J

Epidemiol. 1991;20(1):45-53.

67. Hill S, Sarfati D, Blakely T, Robson B, Purdie G, Chen J, Dennett E, Cormack D,

Cunningham R, Dew K, McCreanor T, Kawachi I. Survival disparities in Indigenous and

non-Indigenous New Zealanders with colon cancer: the role of patient comorbidity,

treatment and health service factors. J Epidemiol Community Health. 2010;64(2):117-

23.

68. Ajwani S, Blakely T, Robson B, Tobias M, Bonne M. Decades of disparity: ethnic

mortality trends in New Zealand 1980-1999 [Internet]. Ministry of Health and

University of Otago, New Zealand; 2003.

69. Robson B, Cormack D. Unequal impact: Māori and non-Māori cancer statistics 1996-

2001. Wellington: Ministry of Health; 2006.

70. McKenzie F, Jeffreys M, t Mannetje A, Pearce N. Prognostic factors in women with

breast cancer: inequalities by ethnicity and socioeconomic position in New Zealand.

Cancer Causes Control. 2008;19(4):403-11.

71. Sterne JA, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, Wood AM,

Carpenter JR. Multiple imputation for missing data in epidemiological and clinical

research: potential and pitfalls. BMJ. 2009;338:b2393.

72. Eisemann N, Waldmann A, Katalinic A. Imputation of missing values of tumour stage in

population-based cancer registration. BMC Med Res Methodol. 2011;11:129.


282

73. Meredith I, Sarfati D, Ikeda T, Blakely T. Cancer in Pacific people in New Zealand.

Cancer Causes Control. 2012;23(7):1173-84.

74. Weston MK, Moss DP, Stewart J, Hill AG. Differences in breast cancer biological

characteristics between ethnic groups in New Zealand. Breast Cancer Res Treat.

2008;111(3):555-8.

75. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin.

2013;63(1):11-30.

76. Damon A. Race, ethnic group, and disease. Biodemography Soc Biol. 1969;16(2):69-80.

77. Boring CC, Squires TS, Tong T. Cancer statistics 1993. CA Cancer J Clin. 1993;43(1):7-

26.

78. Axtell LM, Myers MH. Contrasts in survival of black and white cancer patients, 1960-

73. J Natl Cancer Inst. 1978;60(6):1209-15.

79. Rose DP, Royak-Schaler R. Tumor biology and prognosis in black breast cancer

patients: a review. Cancer Detect Prev. 2001;25(1):16-31.

80. Li CI, Malone KE, Daling JR. Differences in breast cancer stage, treatment, and survival

by race and ethnicity. Arch Intern Med. 2003;163(1):49-56.

81. Silber JH, Rosenbaum PR, Clark AS, Giantonio BJ, Ross RN, Teng Y, Wang M,

Niknam BA, Ludwig JM, Wang W, Even-Shoshan O, Fox KR. Characteristics

associated with differences in survival among black and white women with breast

cancer. JAMA. 2013;310(4):389-97.

82. Newman LA, Griffith KA, Jatoi I, Simon MS, Crowe JP, Colditz GA. Meta-analysis of

survival in African American and white American patients with breast cancer: ethnicity

compared with socioeconomic status. J Clin Oncol. 2006;24(9):1342-9.

83. Curtis E, Quale C, Haggstrom D, Smith-Bindman R. Racial and ethnic differences in

breast cancer survival: how much is explained by screening, tumor severity, biology,

treatment, comorbidities, and demographics? Cancer. 2008;112(1):171-80.

84. Gerend MA, Pai M. Social determinants of Black-White disparities in breast cancer

mortality: a review. Cancer Epidemiol Biomarkers Prev. 2008;17(11):2913-23.

85. Ward E, Jemal A, Cokkinides V, Singh GK, Cardinez C, Ghafoor A, Thun M. Cancer

disparities by race/ethnicity and socioeconomic status. CA Cancer J Clin. 2004;54(2):78-

93.

86. Li CI, Malone KE, Daling JR. Differences in breast cancer hormone receptor status and

histology by race and ethnicity among women 50 years of age and older. Cancer

Epidemiol Biomarkers Prev. 2002;11(7):601-7.

87. Condon JR, Armstrong BK, Barnes A, Cunningham J. Cancer in Indigenous Australians:

a review. Cancer Causes Control. 2003;14(2):109-21.

88. Valery PC, Coory M, Stirling J, Green AC. Cancer diagnosis, treatment, and survival in

Indigenous and non-Indigenous Australians: a matched cohort study. Lancet.

2006;367(9525):1842-8.

References

283

89. Cunningham J, Rumbold AR, Zhang X, Condon JR. Incidence, aetiology, and outcomes

of cancer in Indigenous peoples in Australia. Lancet Oncol. 2008;9(6):585-95.

90. Jack RH, Davies EA, Moller H. Breast cancer incidence, stage, treatment and survival in

ethnic groups in South East England. Br J Cancer. 2009;100(3):545-50.

91. Farooq S, Coleman MP. Breast cancer survival in South Asian women in England and

Wales. J Epidemiol Community Health. 2005;59(5):402-6.

92. dos Santos Silva I, Mangtani P, De Stavola B, Bell J, Quinn M, Mayer D. Survival from

breast cancer among South Asian and non-South Asian women resident in South East

England. Br J Cancer. 2003;89(3):508-12.

93. National Cancer Institute. SEER Cancer Statistics review. USA; 2011. Contract No.: 10-

5581.

94. Stevens W, Stevens G, Kolbe J, Cox B. Comparison of New Zealand Cancer Registry

data with an independent lung cancer audit. N Z Med J. 2008;121(1276):29-41.

95. Walter F, Webster A, Scott S, Emery J. The Andersen Model of Total Patient Delay: a

systematic review of its application in cancer diagnosis. J Health Serv Res Policy.

2012;17(2):110-8.

96. Mandelblatt JS, Yabroff KR, Kerner JF. Equitable access to cancer services: A review of

barriers to quality care. Cancer. 1999;86(11):2378-90.

97. Cormack D, Purdie G, Ratima M. Access for cancer services for Māori, A report

prepared for the Ministry of Health [Internet].2005.

98. Davis P, Lay-Yee R, Dyall L, Briant R, Sporle A, Brunt D, Scott A. Quality of hospital

care for Māori patients in New Zealand: retrospective cross-sectional assessment.

Lancet. 2006;367(9526):1920-5.

99. Crengle S. Primary care and Māori: Findings from the National primary medical care

survey In: Robson B, Harris R, editors. Hauora: Māori Standards of Health IV A study

of the years 2000-2005. Wellington: Te Rōpū Rangahau Hauora a Eru Pōmare; 2007.

100. Ellison-Loschmann L, Pearce N. Improving access to health care among New Zealand's

Māori population. Am J Public Health. 2006;96(4):612-7.

101. Walker T, Signal L, Russell M, Smiler K, Tuhiwai-Ruru R. The road we travel: Māori

experience of cancer. N Z Med J. 2008;121(1279):27-35.

102. Baxter J. Barriers to health care for Mäori with known diabetes: a literature review and

summary of issues Dunedin: Te Röpü Rangahau Hauora a Ngäi Tahu; 2002.

103. Kickbusch I, Wait S, Maag D. Navigating Health: The Role of Health Literacy. London:

Alliance for Health and the Future; 2006.

104. Forum EHP. Recommendations on Health Information Brussels, Belgium: EU Health

policy Forum; 2005.

105. Rudd RE. Needed action in health literacy. J Health Psychol. 2013;18(8):1004-10.


284

106. Ministry of Health. Korero Mārama: Health Literacy and Māori: Results from the 2006

Adult Literacy and Life Skills Survey Wellington; 2010.

107. Richards MA. The National Awareness and Early Diagnosis Initiative in England:

assembling the evidence. Br J Cancer. 2009;101 Suppl 2:S1-4.

108. Gordon NH, Crowe JP, Brumberg DJ, Berger NA. Socioeconomic factors and race in

breast cancer recurrence and survival. Am J Epidemiol. 1992;135(6):609-18.

109. Dayal HH, Power RN, Chiu C. Race and socio-economic status in survival from breast

cancer. J Chronic Dis. 1982;35(8):675-83.

110. Woods LM, Rachet B, Coleman MP. Origins of socio-economic inequalities in cancer

survival: a review. Ann Oncol. 2006;17(1):5-19.

111. Gordon NH. Socioeconomic factors and breast cancer in black and white Americans.

Cancer Metastasis Rev. 2003;22(1):55-65.

112. Thomson CS, Hole DJ, Twelves CJ, Brewster DH, Black RJ. Prognostic factors in

women with breast cancer: distribution by socioeconomic status and effect on

differences in survival. J Epidemiol Community Health. 2001;55(5):308-15.

113. Yu XQ, O'Connell DL, Gibberd RW, Armstrong BK. Assessing the impact of socio-

economic status on cancer survival in New South Wales, Australia 1996-2001. Cancer

Causes Control. 2008;19(10):1383-90.

114. Lagerlund M, Bellocco R, Karlsson P, Tejler G, Lambe M. Socio-economic factors and

breast cancer survival--a population-based cohort study (Sweden). Cancer Causes

Control. 2005;16(4):419-30.

115. Jeffreys M, Sarfati D, Stevanovic V, Tobias M, Lewis C, Pearce N, Blakely T.

Socioeconomic inequalities in cancer survival in New Zealand: the role of extent of

disease at diagnosis. Cancer Epidemiol Biomarkers Prev. 2009;18(3):915-21.

116. LaVeist TA. Disentangling race and socioeconomic status: a key to understanding health

inequalities. J Urban Health. 2005;82(2 Suppl 3):iii26-34.

117. DeSantis C, Jemal A, Ward E. Disparities in breast cancer prognostic factors by race,

insurance status, and education. Cancer Causes Control. 2010;21(9):1445-50.

118. Hall S, Holman CD, Sheiner H, Hendrie D. The influence of socio-economic and

locational disadvantage on survival after a diagnosis of lung or breast cancer in Western

Australia. J Health Serv Res Policy. 2004;9 Suppl 2:10-6.

119. Jong KE, Vale PJ, Armstrong BK. Rural inequalities in cancer care and outcome. Med J

Aust. 2005;182(1):13-4.

120. Liff JM, Chow WH, Greenberg RS. Rural-urban differences in stage at diagnosis.

Possible relationship to cancer screening. Cancer. 1991;67(5):1454-9.

121. Launoy G, Le Coutour X, Gignoux M, Pottier D, Dugleux G. Influence of rural

environment on diagnosis, treatment, and prognosis of colorectal cancer. J Epidemiol

Community Health. 1992;46(4):365-7.

References

285

122. Haynes R, Pearce J, Barnett R. Cancer survival in New Zealand: ethnic, social and

geographical inequalities. Soc Sci Med. 2008;67(6):928-37.

123. Campbell NC, Elliott AM, Sharp L, Ritchie LD, Cassidy J, Little J. Rural factors and

survival from cancer: analysis of Scottish cancer registrations. Br J Cancer.

2000;82(11):1863-6.

124. Menck HR, Mills PK. The influence of urbanization, age, ethnicity, and income on the

early diagnosis of breast carcinoma: opportunity for screening improvement. Cancer.

2001;92(5):1299-304.

125. Amey CH, Miller MK, Albrecht SL. The role of race and residence in determining stage

at diagnosis of breast cancer. J Rural Health. 1997;13(2):99-108.

126. Wilkinson D, Cameron K. Cancer and cancer risk in South Australia: what evidence for

a rural-urban health differential? The Australian journal of rural health. 2004;12(2):61-6.

127. Sturgeon SR, Schairer C, Gail M, McAdams M, Brinton LA, Hoover RN. Geographic

variation in mortality from breast cancer among white women in the United States. J

Natl Cancer Inst. 1995;87(24):1846-53.

128. Armstrong W, Borman B. Breast cancer in New Zealand: trends, patterns, and data

quality. N Z Med J. 1996;109(1024):221-4.

129. Gill AJ, Martin IG. Survival from upper gastrointestinal cancer in New Zealand: The

effect of distance from a major hospital, socio-economic status, ethnicity, age and

gender. ANZ J Surg. 2002;72(9):643-6.

130. Marmot MG, Altman DG, Cameron DA, Dewar JA, Thompson SG, Wilcox M. The

benefits and harms of breast cancer screening: an independent review. Br J Cancer.

2013;108(11):2205-40.

131. Tabar L, Fagerberg CJ, Gad A, Baldetorp L, Holmberg LH, Grontoft O, Ljungquist U,

Lundstrom B, Manson JC, Eklund G, et al. Reduction in mortality from breast cancer

after mass screening with mammography. Randomised trial from the Breast Cancer

Screening Working Group of the Swedish National Board of Health and Welfare.

Lancet. 1985;1(8433):829-32.

132. Page A, Taylor R. BreastScreen Aotearoa: Independant Monitoring Report, Screening

and assessment report of women attending BSA [Internet]. National Screening Unit,

Ministry of Health, Wellington2011.

133. Curtis E, Wright C, Wall M. The epidemiology of breast cancer in Māori women in

Aotearoa New Zealand: implications for ethnicity data analysis. N Z Med J.

2005;118(1209):U1298.

134. Shapiro S, Coleman EA, Broeders M, Codd M, de Koning H, Fracheboud J, Moss S,

Paci E, Stachenko S, Ballard-Barbash R. Breast cancer screening programmes in 22

countries: current policies, administration and guidelines. International Breast Cancer

Screening Network (IBSN) and the European Network of Pilot Projects for Breast

Cancer Screening. Int J Epidemiol. 1998;27(5):735-42.


286

135. van Ravesteyn NT, Schechter CB, Near AM, Heijnsdijk EA, Stoto MA, Draisma G, de

Koning HJ, Mandelblatt JS. Race-specific impact of natural history, mammography

screening, and adjuvant treatment on breast cancer mortality rates in the United States.

Cancer Epidemiol Biomarkers Prev. 2011;20(1):112-22.

136. Jacobellis J, Cutter G. Mammography screening and differences in stage of disease by

race/ethnicity. Am J Public Health. 2002;92(7):1144-50.

137. Aarts MJ, Voogd AC, Duijm LE, Coebergh JW, Louwman WJ. Socioeconomic

inequalities in attending the mass screening for breast cancer in the south of the

Netherlands: associations with stage at diagnosis and survival. Breast Cancer Res Treat.

2011;128(2):517-25.

138. Kim J, Jang SN. Socioeconomic disparities in breast cancer screening among US

women: trends from 2000 to 2005. J Prev Med Public Health. 2008;41(3):186-94.

139. Peek ME, Han JH. Disparities in screening mammography. Current status, interventions

and implications. J Gen Intern Med. 2004;19(2):184-94.

140. Tessaro I, Eng E, Smith J. Breast cancer screening in older African-American women:

qualitative research findings. Am J Health Promot. 1994;8(4):286-92.

141. Thompson HS, Valdimarsdottir HB, Winkel G, Jandorf L, Redd W. The Group-Based

Medical Mistrust Scale: psychometric properties and association with breast cancer

screening. Prev Med. 2004;38(2):209-18.

142. McNoe B, Richardson AK, Elwood JM. Factors affecting participation in

mammography screening. N Z Med J. 1996;109(1030):359-61.

143. Thomson RM, Crengle S, Lawrenson R. Improving participation in breast screening in a

rural general practice with a predominately Māori population. N Z Med J.

2009;122(1291):39-47.

144. Piccirillo JF, Tierney RM, Costas I, Grove L, Spitznagel EL, Jr. Prognostic importance

of comorbidity in a hospital-based cancer registry. JAMA. 2004;291(20):2441-7.

145. Read WL, Tierney RM, Page NC, Costas I, Govindan R, Spitznagel EL, Piccirillo JF.

Differential prognostic impact of comorbidity. J Clin Oncol. 2004;22(15):3099-103.

146. Tammemagi CM, Nerenz D, Neslund-Dudas C, Feldkamp C, Nathanson D. Comorbidity

and survival disparities among black and white patients with breast cancer. JAMA.

2005;294(14):1765-72.

147. Hawfield A, Lovato J, Covington D, Kimmick G. Retrospective study of the effect of

comorbidity on use of adjuvant chemotherapy in older women with breast cancer in a

tertiary care setting. Crit Rev Oncol Hematol. 2006;59(3):250-5.

148. Satariano WA, Ragland DR. The effect of comorbidity on 3-year survival of women

with primary breast cancer. Ann Intern Med. 1994;120(2):104-10.

149. Satariano WA. Aging, comorbidity, and breast cancer survival: an epidemiologic view.

Adv Exp Med Biol. 1993;330:1-11.

References

287

150. Eley JW, Hill HA, Chen VW, Austin DF, Wesley MN, Muss HB, Greenberg RS, Coates

RJ, Correa P, Redmond CK, et al. Racial differences in survival from breast cancer.

Results of the National Cancer Institute Black/White Cancer Survival Study. JAMA.

1994;272(12):947-54.

151. Maskarinec G, Pagano IS, Yamashiro G, Issell BF. Influences of ethnicity, treatment,

and comorbidity on breast cancer survival in Hawaii. J Clin Epidemiol. 2003;56(7):678-

85.

152. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying

prognostic comorbidity in longitudinal studies: development and validation. J Chronic

Dis. 1987;40(5):373-83.

153. West DW, Satariano WA, Ragland DR, Hiatt RA. Comorbidity and breast cancer

survival: a comparison between black and white women. Annals of epidemiology.

1996;6(5):413-9.

154. Sarfati D, Gurney J, Lim BT, Bagheri N, Simpson A, Koea J, Dennett E. Identifying

important comorbidity among cancer populations using administrative data: Prevalence

and impact on survival. Asia Pac J Clin Oncol. 2013.

155. Brewer N, Borman B, Sarfati D, Jeffreys M, Fleming ST, Cheng S, Pearce N. Does

comorbidity explain the ethnic inequalities in cervical cancer survival in New Zealand?

A retrospective cohort study. BMC Cancer. 2011;11:132.

156. Sarfati D, Hill S, Blakely T, Robson B, Purdie G, Dennett E, Cormack D, Dew K. The

effect of comorbidity on the use of adjuvant chemotherapy and survival from colon

cancer: a retrospective cohort study. BMC Cancer. 2009;9:116.

157. Ministry of Health. A portrait of health: Key results of the 2006/07 New Zealand Health

Survey [Internet]. Wellington2008.

158. Carmichael AR. Obesity as a risk factor for development and poor prognosis of breast

cancer. BJOG. 2006;113(10):1160-6.

159. Carmichael AR, Bates T. Obesity and breast cancer: a review of the literature. Breast.

2004;13(2):85-92.

160. Niraula S, Ocana A, Ennis M, Goodwin PJ. Body size and breast cancer prognosis in

relation to hormone receptor and menopausal status: a meta-analysis. Breast Cancer Res

Treat. 2012;134(2):769-81.

161. Chlebowski RT. Obesity and breast cancer outcome: adding to the evidence. J Clin

Oncol. 2012;30(2):126-8.

162. Ng M, Fleming T, Robinson M, Thomson B, Graetz N, Margono C, Mullany EC,

Biryukov S, Abbafati C, Abera SF, Abraham JP, Abu-Rmeileh NM, Achoki T,

AlBuhairan FS, Alemu ZA, Alfonso R, Ali MK, Ali R, Guzman NA, Ammar W, Anwari

P, Banerjee A, Barquera S, Basu S, Bennett DA, Bhutta Z, Blore J, Cabral N, Nonato IC,

Chang JC, Chowdhury R, Courville KJ, Criqui MH, Cundiff DK, Dabhadkar KC,

Dandona L, Davis A, Dayama A, Dharmaratne SD, Ding EL, Durrani AM, Esteghamati

A, Farzadfar F, Fay DF, Feigin VL, Flaxman A, Forouzanfar MH, Goto A, Green MA,


288

Gupta R, Hafezi-Nejad N, Hankey GJ, Harewood HC, Havmoeller R, Hay S, Hernandez

L, Husseini A, Idrisov BT, Ikeda N, Islami F, Jahangir E, Jassal SK, Jee SH, Jeffreys M,

Jonas JB, Kabagambe EK, Khalifa SE, Kengne AP, Khader YS, Khang YH, Kim D,

Kimokoti RW, Kinge JM, Kokubo Y, Kosen S, Kwan G, Lai T, Leinsalu M, Li Y, Liang

X, Liu S, Logroscino G, Lotufo PA, Lu Y, Ma J, Mainoo NK, Mensah GA, Merriman

TR, Mokdad AH, Moschandreas J, Naghavi M, Naheed A, Nand D, Narayan KM,

Nelson EL, Neuhouser ML, Nisar MI, Ohkubo T, Oti SO, Pedroza A, Prabhakaran D,

Roy N, Sampson U, Seo H, Sepanlou SG, Shibuya K, Shiri R, Shiue I, Singh GM, Singh

JA, Skirbekk V, Stapelberg NJ, Sturua L, Sykes BL, Tobias M, Tran BX, Trasande L,

Toyoshima H, van de Vijver S, Vasankari TJ, Veerman JL, Velasquez-Melendez G,

Vlassov VV, Vollset SE, Vos T, Wang C, Wang SX, Weiderpass E, Werdecker A,

Wright JL, Yang YC, Yatsuya H, Yoon J, Yoon SJ, Zhao Y, Zhou M, Zhu S, Lopez AD,

Murray CJ, Gakidou E. Global, regional, and national prevalence of overweight and

obesity in children and adults during 1980-2013: a systematic analysis for the Global

Burden of Disease Study 2013. Lancet. 2014.

163. Kelsey JL, Bernstein L. Epidemiology and prevention of breast cancer. Annu Rev Public

Health. 1996;17:47-67.

164. Jeffreys M, McKenzie F, Firestone R, Gray M, Cheng S, Moala A, Pearce N, Ellison-

Loschmann L. A multi-ethnic breast cancer case-control study in New Zealand: evidence

of differential risk patterns. Cancer Causes Control. 2013;24(1):135-52.

165. Carey LA, Perou CM, Livasy CA, Dressler LG, Cowan D, Conway K, Karaca G,

Troester MA, Tse CK, Edmiston S, Deming SL, Geradts J, Cheang MC, Nielsen TO,

Moorman PG, Earp HS, Millikan RC. Race, breast cancer subtypes, and survival in the

Carolina Breast Cancer Study. JAMA. 2006;295(21):2492-502.

166. O'Brien KM, Cole SR, Tse CK, Perou CM, Carey LA, Foulkes WD, Dressler LG,

Geradts J, Millikan RC. Intrinsic breast tumor subtypes, race, and long-term survival in

the Carolina Breast Cancer Study. Clin Cancer Res. 2010;16(24):6100-10.

167. Jack RH, Davies EA, Renshaw C, Tutt A, Grocock MJ, Coupland VH, Moller H.

Differences in breast cancer hormone receptor status in ethnic groups: a London

population. Eur J Cancer. 2013;49(3):696-702.

168. Chen VW, Correa P, Kurman RJ, Wu XC, Eley JW, Austin D, Muss H, Hunter CP,

Redmond C, Sobhan M, et al. Histological characteristics of breast carcinoma in blacks

and whites. Cancer Epidemiol Biomarkers Prev. 1994;3(2):127-35.

169. Dachs GU, Kano M, Volkova E, Morrin HR, Davey VC, Harris GC, Cheale M,

Frampton C, Currie MJ, Wells JE, Robinson BA. A profile of prognostic and molecular

factors in European and Māori breast cancer patients. BMC Cancer. 2010;10:543.

170. Chavez-Macgregor M, Litton J, Chen H, Giordano SH, Hudis CA, Wolff AC, Valero V,

Hortobagyi GN, Bondy ML, Gonzalez-Angulo AM. Pathologic complete response in

breast cancer patients receiving anthracycline- and taxane-based neoadjuvant

chemotherapy: evaluating the effect of race/ethnicity. Cancer. 2010;116(17):4168-77.

References

289

171. Fisher B, Redmond C, Fisher ER, Caplan R. Relative worth of estrogen or progesterone

receptor and pathologic characteristics of differentiation as indicators of prognosis in

node negative breast cancer patients: findings from National Surgical Adjuvant Breast

and Bowel Project Protocol B-06. J Clin Oncol. 1988;6(7):1076-87.

172. Onitilo AA, Engel JM, Greenlee RT, Mukesh BN. Breast cancer subtypes based on

ER/PR and Her2 expression: comparison of clinicopathologic features and survival. Clin

Med Res. 2009;7(1-2):4-13.

173. Mason BH, Holdaway IM, Mullins PR, Yee LH, Kay RG. Progesterone and estrogen

receptors as prognostic variables in breast cancer. Cancer Res. 1983;43(6):2985-90.

174. Kurian AW, Fish K, Shema SJ, Clarke CA. Lifetime risks of specific breast cancer

subtypes among women in four racial/ethnic groups. Breast Cancer Res.

2010;12(6):R99.

175. Simmons D, Weblemoe T, Voyle J, Prichard A, Leakehe L, Gatland B. Personal barriers

to diabetes care: lessons from a multi-ethnic community in New Zealand. Diabet Med.

1998;15(11):958-64.

176. Westbrooke I, Baxter J, Hogan J. Are Māori under-served for cardiac interventions? N Z

Med J. 2001;114(1143):484-7.

177. Curtis E. Deserving of more: framing of Māori inequities in cardiovascular care remains

a challenge. N Z Med J. 2013;126(1379).

178. Kerr A, Mustafa A, Lee M, Wells S, Grey C, Riddell T, Harrison W. Ethnicity and

revascularisation following acute coronary syndromes: a 5-year cohort study (ANZACS-

QI-3). N Z Med J. 2013;127(1393):38-51.

179. Harris R, Robson B, Curtis E, Purdie G, Cormack D, Reid P. Māori and non-Māori

differences in caesarean section rates: a national review. N Z Med J. 2007;120(1250).

180. McLeod D, Morgan S, McKinlay E, Dew K, Cumming J, Dowell A, Love T. Use of, and

attitudes to, clinical priority assessment criteria in elective surgery in New Zealand. J

Health Serv Res Policy. 2004;9(2):91-9.

181. Hill S, Sarfati D, Blakely T, Robson B, Purdie G, Dennett E, Cormack D, Dew K,

Ayanian JZ, Kawachi I. Ethnicity and management of colon cancer in New Zealand: do

indigenous patients get a worse deal? Cancer. 2010;116(13):3205-14.

182. Stevens W, Stevens G, Kolbe J, Cox B. Ethnic differences in the management of lung

cancer in New Zealand. J Thorac Oncol. 2008;3(3):237-44.

183. Tukuitonga C, Bindman A. Ethnic and gender differences in the use of coronary artery

revascularisation procedures in New Zealand. N Z Med J. 2002;115(1152):179-82.

184. Hill S, Sarfati D, Robson B, Blakely T. Indigenous inequalities in cancer: what role for

health care? ANZ J Surg. 2013;83(1-2):36-41.

185. Bach PB, Cramer LD, Warren JL, Begg CB. Racial differences in the treatment of early-

stage lung cancer. N Engl J Med. 1999;341(16):1198-205.


290

186. Steyerberg EW, Earle CC, Neville BA, Weeks JC. Racial differences in surgical

evaluation, treatment, and outcome of locoregional esophageal cancer: a population-

based analysis of elderly patients. J Clin Oncol. 2005;23(3):510-7.

187. Tewari A, Horninger W, Pelzer AE, Demers R, Crawford ED, Gamito EJ, Divine G,

Johnson CC, Bartsch G, Menon M. Factors contributing to the racial differences in

prostate cancer mortality. BJU Int. 2005;96(9):1247-52.

188. Sloane D, Chen H, Howell C. Racial disparity in primary hepatocellular carcinoma:

tumor stage at presentation, surgical treatment and survival. J Natl Med Assoc.

2006;98(12):1934-9.

189. Randall TC, Armstrong K. Differences in treatment and outcome between African-

American and white women with endometrial cancer. J Clin Oncol. 2003;21(22):4200-6.

190. Ball JK, Elixhauser A. Treatment differences between blacks and whites with colorectal

cancer. Med Care. 1996;34(9):970-84.

191. Shavers VL, Brown ML. Racial and ethnic disparities in the receipt of cancer treatment.

J Natl Cancer Inst. 2002;94(5):334-57.

192. Morris CR, Cohen R, Schlag R, Wright WE. Increasing trends in the use of breast-

conserving surgery in California. Am J Public Health. 2000;90(2):281-4.

193. Mandelblatt JS, Hadley J, Kerner JF, Schulman KA, Gold K, Dunmore-Griffith J, Edge

S, Guadagnoli E, Lynch JJ, Meropol NJ, Weeks JC, Winn R. Patterns of breast

carcinoma treatment in older women: patient preference and clinical and physical

influences. Cancer. 2000;89(3):561-73.

194. Tropman SE, Ricketts TC, Paskett E, Hatzell TA, Cooper MR, Aldrich T. Rural breast

cancer treatment: evidence from the Reaching Communities for Cancer Care (REACH)

project. Breast Cancer Res Treat. 1999;56(1):59-66.

195. Siminoff LA, Zhang A, Saunders Sturm CM, Colabianchi N. Referral of breast cancer

patients to medical oncologists after initial surgical management. Med Care.

2000;38(7):696-704.

196. Kalager M, Haldorsen T, Bretthauer M, Hoff G, Thoresen SO, Adami HO. Improved

breast cancer survival following introduction of an organized mammography screening

program among both screened and unscreened women: a population-based cohort study.

Breast Cancer Res. 2009;11(4):R44.

197. Page A, Taylor R. BreastScreen Aotearoa: Independent Monitoring Report -Treatment

of women with BSA detected cancers (women screened July 2005-June 2007).

Wellington; 2009.

198. Richards MA, Westcombe AM, Love SB, Littlejohns P, Ramirez AJ. Influence of delay

on survival in patients with breast cancer: a systematic review. Lancet.

1999;353(9159):1119-26.

199. Unger-Saldana K, Infante-Castaneda C. Delay of medical care for symptomatic breast

cancer: a literature review. Salud Publica Mex. 2009;51 Suppl 2:s270-85.

References

291

200. Meechan G, Collins J, Petrie K. Delay in seeking medical care for self-detected breast

symptoms in New Zealand women. N Z Med J. 2002;115(1166):U257.

201. Olesen F, Hansen RP, Vedsted P. Delay in diagnosis: the experience in Denmark. Br J

Cancer. 2009;101 Suppl 2:S5-8.

202. Facione NC. Delay versus help seeking for breast cancer symptoms: a critical review of

the literature on patient and provider delay. Soc Sci Med. 1993;36(12):1521-34.

203. Nosarti C, Crayford T, Roberts JV, Elias E, McKenzie K, David AS. Delay in

presentation of symptomatic referrals to a breast clinic: patient and system factors. Br J

Cancer. 2000;82(3):742-8.

204. Caplan LS, May DS, Richardson LC. Time to diagnosis and treatment of breast cancer:

results from the National Breast and Cervical Cancer Early Detection Program, 1991-

1995. Am J Public Health. 2000;90(1):130-4.

205. Richards MA, Westcombe AM, Ramirez A J, Love SB, Littlejohns P. A systematic

review of the delay in diagnosis/treatment of symptomatic breast cancer: a review

commissioned by the NHS Cancer Research and development Programme

[Internet].1999.

206. McLaughlin JM, Anderson RT, Ferketich AK, Seiber EE, Balkrishnan R, Paskett ED.

Effect on survival of longer intervals between confirmed diagnosis and treatment

initiation among low-income women with breast cancer. J Clin Oncol.

2012;30(36):4493-500.

207. Biagi JJ, Raphael M, King WD, Kong W, Booth CM, Mackillop WJ. The effect of delay

in time to adjuvant chemotherapy (TTAC) on survival in breast cancer (BC): a

systematic review and meta-analysis J Clin Oncol. 2011;29(15_suppl, 2011 ASCO

Annual Meeting Abstracts ):1128.

208. Yu KD, Huang S, Zhang JX, Liu GY, Shao ZM. Association between delayed initiation

of adjuvant CMF or anthracycline-based chemotherapy and survival in breast cancer: a

systematic review and meta-analysis. BMC Cancer. 2013;13:240.

209. Hershman DL, Wang X, McBride R, Jacobson JS, Grann VR, Neugut AI. Delay in

initiating adjuvant radiotherapy following breast conservation surgery and its impact on

survival. Int J Radiat Oncol Biol Phys. 2006;65(5):1353-60.

210. Gagliato Dde M, Gonzalez-Angulo AM, Lei X, Theriault RL, Giordano SH, Valero V,

Hortobagyi GN, Chavez-Macgregor M. Clinical impact of delaying initiation of adjuvant

chemotherapy in patients with breast cancer. J Clin Oncol. 2014;32(8):735-44.

211. Pomare E, Tutengaehe H, Ramsden I, Hight M, Pearce N, Ormsby V. He Mate Huango:

Māori Asthma Review. Wellington; 1991.

212. Howard DL, Penchansky R, Brown MB. Disaggregating the effects of race on breast

cancer survival. Fam Med. 1998;30(3):228-35.

213. Bhatta S, Hou N, Huo D, Polite B, Fleming G, Olopade O, Hong S, editors. Compliance

to adjuvant hormone therapy for black and white women with breast cancer. ASCO

Annual Meeting 2011.


292

214. Shahid S, Thompson SC. An overview of cancer and beliefs about the disease in

Indigenous people of Australia, Canada, New Zealand and the US. Aust N Z J Public

Health. 2009;33(2):109-18.

215. Wilson D. How Māori women experience 'mainstream' health services. Nurs N Z. 2008

July.

216. Early Breast Cancer Trialists' Collaborative Group (EBCTCG). Effects of chemotherapy

and hormonal therapy for early breast cancer on recurrence and 15-year survival: an

overview of the randomised trials. Lancet. 2005;365(9472):1687-717.

217. Hershman DL, Shao T, Kushi LH, Buono D, Tsai WY, Fehrenbacher L, Kwan M,

Gomez SL, Neugut AI. Early discontinuation and non-adherence to adjuvant hormonal

therapy are associated with increased mortality in women with breast cancer. Breast

Cancer Res Treat. 2011;126(2):529-37.

218. Ooi CW, Campbell ID, Kollias J, de Silva P. National Breast Cancer Audit: overview of

invasive breast cancer in New Zealand. N Z Med J. 2012;125(1359):7-16.

219. Chlebowski RT, Geller ML. Adherence to endocrine therapy for breast cancer.

Oncology. 2006;71(1-2):1-9.

220. Hadji P, Ziller V, Kyvernitakis J, Bauer M, Haas G, Schmidt N, Kostev K. Persistence in

patients with breast cancer treated with tamoxifen or aromatase inhibitors: a

retrospective database analysis. Breast Cancer Res Treat. 2013;138(1):185-91.

221. Kimmick G, Anderson R, Camacho F, Bhosle M, Hwang W, Balkrishnan R. Adjuvant

hormonal therapy use among insured, low-income women with breast cancer. J Clin

Oncol. 2009;27(21):3445-51.

222. Atkins L, Fallowfield L. Intentional and non-intentional non-adherence to medication

amongst breast cancer patients. Eur J Cancer. 2006;42(14):2271-6.

223. Kesson EM, Allardice GM, George WD, Burns HJ, Morrison DS. Effects of

multidisciplinary team working on breast cancer survival: retrospective, comparative,

interventional cohort study of 13 722 women. BMJ. 2012;344:e2718.

224. Whop LJ, Valery PC, Beesley VL, Moore SP, Lokuge K, Jacka C, Garvey G.

Navigating the cancer journey: a review of patient navigator programs for Indigenous

cancer patients. Asia Pac J Clin Oncol. 2012;8(4):e89-96.

225. Hunnibell LS, Rose MG, Connery DM, Grens CE, Hampel JM, Rosa M, Vogel DC.

Using nurse navigation to improve timeliness of lung cancer care at a veterans hospital.

Clin J Oncol Nurs. 2012;16(1):29-36.

226. Haideri NA, Moormeier JA. Impact of patient navigation from diagnosis to treatment in

an urban safety net breast cancer population. J Cancer. 2011;2:467-73.

227. Ferrante JM, Chen PH, Kim S. The effect of patient navigation on time to diagnosis,

anxiety, and satisfaction in urban minority women with abnormal mammograms: a

randomized controlled trial. J Urban Health. 2008;85(1):114-24.

228. Natale-Pereira A, Enard KR, Nevarez L, Jones LA. The role of patient navigators in

eliminating health disparities. Cancer. 2011;117(15 Suppl):3543-52.

References

293

229. Collinson L, Foster RH, Stapleton M, Blakely T. Cancer care coordinators: what are

they and what will they cost? N Z Med J. 2013;126(1381):75-86.

230. Harris R, Cormack D, Tobias M, Yeh LC, Talamaivao N, Minster J, Timutimu R. Self-

Reported Experience of Racial Discrimination and Health Care Use in New Zealand:

Results From the 2006/07 New Zealand Health Survey. Am J Public Health. 2012.

231. Smedley BD, Stith AY, Nelson AR. Unequal treatment: confronting racial and ethnic

disparities in health care. National Academies Press; 2009.

232. Blendon RJ, Altman DE, Benson JM, Brodie M, Buhr T, Deane C, Buscho S. Voters and

health reform in the 2008 presidential election. N Engl J Med. 2008;359(19):2050-61.

233. Barnett JR. Geographic perspectives on hospital restructuring and its impacts in New

Zealand. N Z Med J. 2000;113(1107):125-8.

234. Blustein J. Who is accountable for racial equity in health care? JAMA. 2008;299(7):814-

6.

235. Cumming J, Stillman S, Poland M. Who pays what for primary health care? : Patterns

and determinants of the fees paid by patients in a mixed public-private financing model.

Wellington: Motu Economic and Public Policy Research; 2008.

236. Kaluzny AD. Prevention and control research within a changing health care system. Prev

Med. 1997;26(5):S31-S5.

237. Kenny L, Lehman M. Sequential audits of unacceptable delays in radiation therapy in

Australia and New Zealand. Australas Radiol. 2004;48(1):29-34.

238. Harris R, Tobias M, Jeffreys M, Waldegrave K, Karlsen S, Nazroo J. Racism and health:

the relationship between experience of racial discrimination and health in New Zealand.

Soc Sci Med. 2006;63(6):1428-41.

239. Vernon SW, Laville EA, Jackson GL. Participation in breast screening programs: a

review. Soc Sci Med. 1990;30(10):1107-18.

240. Van Ryn M, Burke J. The effect of patient race and socio-economic status on physicians'

perceptions of patients. Soc Sci Med. 2000;50(6):813-28.

241. Einbinder LC, Schulman KA. The effect of race on the referral process for invasive

cardiac procedures. Med Care Res Rev. 2000;57(4 suppl):162-80.

242. Schulman KA, Berlin JA, Harless W, Kerner JF, Sistrunk S, Gersh BJ, Dube R,

Taleghani CK, Burke JE, Williams S. The effect of race and sex on physicians'

recommendations for cardiac catheterization. N Engl J Med. 1999;340(8):618-26.

243. McCreanor T, Naim R. Tauiwi general practitioners' talk about Māori health:

Interpretative repertoires. N Z Med J. 2002;115(1167):U272.

244. Fox SA, Stein JA. The effect of physician-patient communication on mammography

utilization by different ethnic groups. Med Care. 1991;29(11):1065-82.

245. Waitzkin H. Information giving in medical care. Journal of health and social behavior.

1985;26(2):81-101.


294

246. Pendleton DA, Bochner S. The communication of medical information in general

practice consultations as a function of patients' social class. Social science & medicine

Medical psychology & medical sociology. 1980;14A(6):669-73.

247. el-Sadr W, Capps L. The challenge of minority recruitment in clinical trials for AIDS.

JAMA. 1992;267(7):954-7.

248. Kerr S, Penney L, Barnes HM, McCreanor T. Kaupapa Māori Action Research to

improve heart disease services in Aotearoa, New Zealand. Ethn Health. 2010;15(1):15-

31.

249. Kidd J, Gibbons V, Kara E, Blundell R, Berryman K. A whanau ora journey of Māori

men with chronic illness: A Te Korowai analysis. AlterNative: An International Journal

of Indigenous Peoples. 2013;Vol. 9(No. 2):125-41.

250. Eckermann A-K. Binan Goonj : bridging cultures in Aboriginal health. 3rd ed. Sydney:

Churchill Livingstone/Elsevier; 2010.

251. Williams R. Cultural safety: what does it mean for our work practice? Aust N Z J Public

Health. 1999;23(2):213-4.

252. Klepin HD, Pitcher BN, Ballman KV, Kornblith AB, Hurria A, Winer EP, Hudis C,

Cohen HJ, Muss HB, Kimmick GG. Comorbidity, Chemotherapy Toxicity, and

Outcomes Among Older Women Receiving Adjuvant Chemotherapy for Breast Cancer

on a Clinical Trial: CALGB 49907 and CALGB 361004 (Alliance). Journal of oncology

practice / American Society of Clinical Oncology. 2014:JOP. 2014.001388.

253. Breen N, Kessler LG, Brown ML. Breast cancer control among the underserve: an

overview. Breast Cancer Res Treat. 1996;40(1):105-15.

254. Schwartz JS, Lewis CE, Clancy C, Kinosian MS, Radany MH, Koplan JP. Internists'

practices in health promotion and disease prevention. A survey. Ann Intern Med.

1991;114(1):46-53.

255. Merrill R, Merrill A, Mayer L. Factors associated with no surgery or radiation therapy

for invasive cervical cancer in Black and White women. Ethn Dis. 1999;10(2):248-56.

256. Cancer Registry Act Stat. Public Act 1993 No 102 (1993) [cited 18/02/2014]. Available

from: http://www.legislation.govt.nz/act/public/1993/0102/latest/DLM318888.html

257. Mortality Collection. Data dictionary, Version 1.3. Wellington: Ministry of Health;

2009.

258. National Centre for Classification in Health. ICD-10: International statistical

classification of diseases and related problems 2nd ed.: World Health Organization;

1999.

259. Neave L, Harvey V, Benjamin C, Thompson P, Pellett O, Whitlock J, Jones W, Poole G.

The Auckland Breast Cancer Register: a special project of the Auckland Breast Cancer

Study Group. N Z Med J. 2003;116(1184):U648.

260. Davey V, Robinson B, Dijkstra B, Harris G. The Christchurch Breast Cancer Patient

Register: the first year. N Z Med J. 2012;125(1360):37-47.

http://www.legislation.govt.nz/act/public/1993/0102/latest/DLM318888.html

References

295

261. IBM Corp. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY2013.

262. American Anthroplogy Association. AAA Statement on Race. Am Anthropol.

1998;100(3):712-3.

263. Parra FC, Amado RC, Lambertucci JR, Rocha J, Antunes CM, Pena SD. Color and

genomic ancestry in Brazilians. Proc Natl Acad Sci U S A. 2003;100(1):177-82.

264. Statistics New Zealand. Report of the review of the measurement of ethnicity.

Wellington: Statistics New Zealand; 2004.

265. Ministry of Health. Ethnicity Data Protocols for the Health and Disability Sector

Wellington 2004 [15/05/2014]. Available from:

http://www.health.govt.nz/publication/ethnicity-data-protocols-health-and-disability-

sector

266. Adamo MB, Johnson CH, Ruhl JL, Dickie LA. SEER Program Coding and Staging

Manual. Bethesda, MD: National Cancer Institute, NIH Publication number 10-5581;

2010

267. American Joint Committee on Cancer. AJCC Cancer Staging Manual. New York:

Springer-Verlag; 2010.

268. Ministry of Health. New Zealand Cancer Registry Wellington 2014 [cited 2014 3rd

Mar]. Available from: http://www.health.govt.nz/nz-health-statistics/national-

collections-and-surveys/collections/new-zealand-cancer-registry-nzcr

269. Elston CW, Ellis IO. Pathological prognostic factors in breast cancer: the value of

histological grade in breast cancer: experience from a large study with long-term follow-

up. Histopathology. 1991;19(5):403-10.

270. Pauletti G, Godolphin W, Press MF, Slamon DJ. Detection and quantitation of HER-

2/neu gene amplification in human breast cancer archival material using fluorescence in

situ hybridization. Oncogene. 1996;13(1):63-72.

271. Salmond C, Crampton P, Atkinson J. NZDep2006: Index of Deprivation. Wellington:

Department of Public Health, University of Otago; 2007.

272. National Breast Cancer Centre Early Breast Cancer Working Group. Clinical practice

guidelines for the management of early breast cancer. Australia: National Health and

Medical Research Council, Australia; 1995.

273. Ministry of Health. Cancer Patient Survival Change Over Time Update: Covering the

Period 1994 to 2009 Wellington; 2012.

274. World Health Organization. National cancer control programmes: policies and

managerial guidelines. Second ed. Geneva2002.

275. New Zealand Cancer Control Trust. New Zealand Cancer Control Strategy [Internet].

Wellington2003. Available from:

http://www.health.govt.nz/system/files/documents/publications/cancercontrolstrategy.pd

f

http://www.health.govt.nz/publication/ethnicity-data-protocols-health-and-disability-sector

http://www.health.govt.nz/publication/ethnicity-data-protocols-health-and-disability-sector

http://www.health.govt.nz/nz-health-statistics/national-collections-and-surveys/collections/new-zealand-cancer-registry-nzcr

http://www.health.govt.nz/nz-health-statistics/national-collections-and-surveys/collections/new-zealand-cancer-registry-nzcr

http://www.health.govt.nz/system/files/documents/publications/cancercontrolstrategy.pdf

http://www.health.govt.nz/system/files/documents/publications/cancercontrolstrategy.pdf


296

276. Skeet RG. Quality and quality control. In: Jensen OM, Parkin DM, MacLennan R, Muir

CS, Skeet RG, editors. Cancer Registration: Principles and Methods. Lyon, France:

International Agency for Research on Cancer; 1991. p. 101-7.

277. Dockerty JD, Becroft DM, Lewis ME, Williams SM. The accuracy and completeness of

childhood cancer registration in New Zealand. Cancer Causes Control. 1997;8(6):857-

64.

278. Gurney J, Sarfati D, Stanley J, Dennett E, Johnson C, Koea J, Simpson A, Studd R.

Unstaged cancer in a population-based registry: prevalence, predictors and patient

prognosis. Cancer Epidemiol. 2013;37(4):498-504.

279. Ording AG, Nielsson MS, Froslev T, Friis S, Garne JP, Sogaard M. Completeness of

breast cancer staging in the Danish Cancer Registry, 2004-2009. Clin Epidemiol. 2012;4

Suppl 2:11-6.

280. Sogaard M, Olsen M. Quality of cancer registry data: completeness of TNM staging and

potential implications. Clin Epidemiol. 2012;4 Suppl 2:1-3.

281. Thoburn KK, German RR, Lewis M, Nichols PJ, Ahmed F, Jackson-Thompson J. Case

completeness and data accuracy in the Centers for Disease Control and Prevention's

National Program of Cancer Registries. Cancer. 2007;109(8):1607-16.

282. Gulliford MC, Bell J, Bourne HM, Petruckevitch A. The reliability of Cancer Registry

records. Br J Cancer. 1993;67(4):819-21.

283. Schouten LJ, Langendijk JA, Jager JJ, van den Brandt PA. Validity of the stage of lung

cancer in records of the Maastricht cancer registry, The Netherlands. Lung Cancer.

1997;17(1):115-22.

284. Dregan A, Moller H, Murray-Thomas T, Gulliford MC. Validity of cancer diagnosis in a

primary care database compared with linked cancer registrations in England. Population-

based cohort study. Cancer Epidemiol. 2012;36(5):425-9.

285. Koroukian SM, Xu F, Beaird H, Diaz M, Murray P, Rose JH. Complexity of care needs

and unstaged cancer in elders: a population-based study. Cancer Detect Prev.

2007;31(3):199-206.

286. Worthington JL, Koroukian SM, Cooper GS. Examining the characteristics of unstaged

colon and rectal cancer cases. Cancer Detect Prev. 2008;32(3):251-8.

287. Lengerich EJ, Tucker TC, Powell RK, Colsher P, Lehman E, Ward AJ, Siedlecki JC,

Wyatt SW. Cancer incidence in Kentucky, Pennsylvania, and West Virginia: disparities

in Appalachia. J Rural Health. 2005;21(1):39-47.

288. Cancer Control New Zealand. Review of the New Zealand Cancer Registry Wellington

2010 Available from:

http://cancercontrolnz.govt.nz/sites/default/files/Review%20of%20the%20NZ%20Cance

r%20Registry.pdf

289. Romond EH, Perez EA, Bryant J, Suman VJ, Geyer CE, Jr., Davidson NE, Tan-Chiu E,

Martino S, Paik S, Kaufman PA, Swain SM, Pisansky TM, Fehrenbacher L, Kutteh LA,

Vogel VG, Visscher DW, Yothers G, Jenkins RB, Brown AM, Dakhil SR, Mamounas

http://cancercontrolnz.govt.nz/sites/default/files/Review%20of%20the%20NZ%20Cancer%20Registry.pdf

http://cancercontrolnz.govt.nz/sites/default/files/Review%20of%20the%20NZ%20Cancer%20Registry.pdf

References

297

EP, Lingle WL, Klein PM, Ingle JN, Wolmark N. Trastuzumab plus adjuvant

chemotherapy for operable HER2-positive breast cancer. N Engl J Med.

2005;353(16):1673-84.

290. Moodie P, Metcalfe S, McNee W. Response from PHARMAC: difficult choices. N Z

Med J. 2003;116(1170):U361.

291. Tabar L, Yen MF, Vitak B, Chen HH, Smith RA, Duffy SW. Mammography service

screening and mortality in breast cancer patients: 20-year follow-up before and after

introduction of screening. Lancet. 2003;361(9367):1405-10.

292. Tabar L, Fagerberg G, Chen HH, Duffy SW, Smart CR, Gad A, Smith RA. Efficacy of

breast cancer screening by age. New results from the Swedish Two-County Trial.

Cancer. 1995;75(10):2507-17.

293. Lantz PM, Mujahid M, Schwartz K, Janz NK, Fagerlin A, Salem B, Liu L, Deapen D,

Katz SJ. The influence of race, ethnicity, and individual socioeconomic factors on breast

cancer stage at diagnosis. Am J Public Health. 2006;96(12):2173-8.

294. Fredholm H, Eaker S, Frisell J, Holmberg L, Fredriksson I, Lindman H. Breast cancer in

young women: poor survival despite intensive treatment. PLoS One. 2009;4(11):e7695.

295. Haakinson DJ, Leeds SG, Dueck AC, Gray RJ, Wasif N, Stucky CC, Northfelt DW,

Apsey HA, Pockaj B. The impact of obesity on breast cancer: a retrospective review.

Ann Surg Oncol. 2012;19(9):3012-8.

296. Jones BA, Kasi SV, Curnen MG, Owens PH, Dubrow R. Severe obesity as an

explanatory factor for the black/white difference in stage at diagnosis of breast cancer.

Am J Epidemiol. 1997;146(5):394-404.

297. Davies EA, Renshaw C, Dixon S, Moller H, Coupland VH. Socioeconomic and ethnic

inequalities in screen-detected breast cancer in London. J Public Health (Oxf).

2013;35(4):607-15.

298. Lagerlund M, Hedin A, Sparen P, Thurfjell E, Lambe M. Attitudes, beliefs, and

knowledge as predictors of nonattendance in a Swedish population-based mammography

screening program. Prev Med. 2000;31(4):417-28.

299. Vona-Davis L, Rose DP. The influence of socioeconomic disparities on breast cancer

tumor biology and prognosis: a review. J Womens Health (Larchmt). 2009;18(6):883-

93.

300. Taylor A, Cheng KK. Social deprivation and breast cancer. J Public Health Med.

2003;25(3):228-33.

301. Amend K, Hicks D, Ambrosone CB. Breast cancer in African-American women:

differences in tumor biology from European-American women. Cancer Res.

2006;66(17):8327-30.

302. Dunn BK, Agurs-Collins T, Browne D, Lubet R, Johnson KA. Health disparities in

breast cancer: biology meets socioeconomic status. Breast Cancer Res Treat.

2010;121(2):281-92.


298

303. Tishkoff SA, Kidd KK. Implications of biogeography of human populations for'race'and

medicine. Nat Genet. 2004;36:S21-S7.

304. Wiencke JK. Impact of race/ethnicity on molecular pathways in human cancer. Nat Rev

Cancer. 2004;4(1):79-84.

305. Allred DC. Commentary: hormone receptor testing in breast cancer: a distress signal

from Canada. Oncologist. 2008;13(11):1134-6.

306. Saint-Jacques N, Younis T, Dewar R, Rayson D. Wait times for breast cancer care. Br J

Cancer. 2007;96(1):162-8.

307. Elmore JG, Nakano CY, Linden HM, Reisch LM, Ayanian JZ, Larson EB. Racial

inequities in the timing of breast cancer detection, diagnosis, and initiation of treatment.

Med Care. 2005;43(2):141-8.

308. Gorin SS, Heck JE, Cheng B, Smith SJ. Delays in breast cancer diagnosis and treatment

by racial/ethnic group. Arch Intern Med. 2006;166(20):2244-52.

309. Wright GP, Wong JH, Morgan JW, Roy-Chowdhury S, Kazanjian K, Lum SS. Time

from diagnosis to surgical treatment of breast cancer: factors influencing delays in

initiating treatment. Am Surg. 2010;76(10):1119-22.

310. Ministry of Health. Cancer: New Registrations and Deaths 2008. Wellington; 2011.

311. Bayard JM, Calianno C, Mee CL. Care coordinator: blending roles to improve patient

outcomes. Nurs Manage. 1997;28(8):49-51; quiz 2.

312. Martin MA, Meyricke R, O'Neill T, Roberts S. Mastectomy or breast conserving

surgery? Factors affecting type of surgical treatment for breast cancer--a classification

tree approach. BMC Cancer. 2006;6:98.

313. Gwyn K, Bondy ML, Cohen DS, Lund MJ, Liff JM, Flagg EW, Brinton LA, Eley JW,

Coates RJ. Racial differences in diagnosis, treatment, and clinical delays in a population-

based study of patients with newly diagnosed breast carcinoma. Cancer.

2004;100(8):1595-604.

314. Pikarsky AJ, Saida Y, Yamaguchi T, Martinez S, Chen W, Weiss EG, Nogueras JJ,

Wexner SD. Is obesity a high-risk factor for laparoscopic colorectal surgery? Surg

Endosc. 2002;16(5):855-8.

315. El-Tamer MB, Ward BM, Schifftner T, Neumayer L, Khuri S, Henderson W. Morbidity

and mortality following breast cancer surgery in women: national benchmarks for

standards of care. Ann Surg. 2007;245(5):665-71.

316. Bickell NA, Wang JJ, Oluwole S, Schrag D, Godfrey H, Hiotis K, Mendez J, Guth AA.

Missed opportunities: racial disparities in adjuvant breast cancer treatment. J Clin Oncol.

2006;24(9):1357-62.

317. Griggs JJ, Culakova E, Sorbero ME, Poniewierski MS, Wolff DA, Crawford J, Dale DC,

Lyman GH. Social and racial differences in selection of breast cancer adjuvant

chemotherapy regimens. J Clin Oncol. 2007;25(18):2522-7.

References

299

318. Lund MJ, Brawley OP, Ward KC, Young JL, Gabram SS, Eley JW. Parity and disparity

in first course treatment of invasive breast cancer. Breast Cancer Res Treat.

2008;109(3):545-57.

319. Fedewa SA, Ward EM, Stewart AK, Edge SB. Delays in adjuvant chemotherapy

treatment among patients with breast cancer are more likely in African American and

Hispanic populations: a national cohort study 2004-2006. J Clin Oncol.

2010;28(27):4135-41.

320. Hershman DL, Wang X, McBride R, Jacobson JS, Grann VR, Neugut AI. Delay of

adjuvant chemotherapy initiation following breast cancer surgery among elderly women.

Breast Cancer Res Treat. 2006;99(3):313-21.

321. Berry DA, Cronin KA, Plevritis SK, Fryback DG, Clarke L, Zelen M, Mandelblatt JS,

Yakovlev AY, Habbema JD, Feuer EJ. Effect of screening and adjuvant therapy on

mortality from breast cancer. N Engl J Med. 2005;353(17):1784-92.

322. Freedman RA, He Y, Winer EP, Keating NL. Racial/Ethnic differences in receipt of

timely adjuvant therapy for older women with breast cancer: are delays influenced by

the hospitals where patients obtain surgical care? Health Serv Res. 2013;48(5):1669-83.

323. Lohrisch C, Paltiel C, Gelmon K, Speers C, Taylor S, Barnett J, Olivotto IA. Impact on

survival of time from definitive surgery to initiation of adjuvant chemotherapy for early-

stage breast cancer. J Clin Oncol. 2006;24(30):4888-94.

324. Colleoni M, Bonetti M, Coates AS, Castiglione-Gertsch M, Gelber RD, Price K,

Rudenstam CM, Lindtner J, Collins J, Thurlimann B, Holmberg S, Veronesi A, Marini

G, Goldhirsch A. Early start of adjuvant chemotherapy may improve treatment outcome

for premenopausal breast cancer patients with tumors not expressing estrogen receptors.

J Clin Oncol. 2000;18(3):584-90.

325. Guidry JJ, Aday LA, Zhang D, Winn RJ. Transportation as a barrier to cancer treatment.

Cancer Pract. 1997;5(6):361-6.

326. Nelson AR. Unequal treatment: report of the Institute of Medicine on racial and ethnic

disparities in healthcare. The Annals of thoracic surgery. 2003;76(4):S1377-81.

327. Mundt AJ, Connell PP, Campbell T, Hwang JH, Rotmensch J, Waggoner S. Race and

clinical outcome in patients with carcinoma of the uterine cervix treated with radiation

therapy. Gynecologic oncology. 1998;71(2):151-8.

328. Rippy EE, Ainsworth R, Sathananthan D, Kollias J, Bochner M, Whitfield R. Influences

on decision for mastectomy in patients eligible for breast conserving surgery. Breast.

2014;23(3):273-8.

329. Schwartz F, Evans W, Sullivan T, Angus H. Gaining access to appropriate cancer

services: a four point strategy to reduce waiting times in Ontario. Ontario, Canada:

Cancer quality council of Ontario, Canada; 2004.

330. Garcia SF, Hahn EA, Jacobs EA. Addressing low literacy and health literacy in clinical

oncology practice. The journal of supportive oncology. 2010;8(2):64-9.


300

331. Woynar S.P, Burban P, Le Prisé E, Romestaing P, Maylin C, Mazeau V, Cauterman M,

Vendrely V. Reducing radiotherapy delays after surgery for breast cancer in five

radiotherapy departments J Clin Oncol. 2007;Vol. 25 (no. 18_suppl 17011 ).

332. Kallen MA, Terrell JA, Lewis-Patterson P, Hwang JP. Improving wait time for

chemotherapy in an outpatient clinic at a comprehensive cancer center. Journal of

oncology practice / American Society of Clinical Oncology. 2012;8(1):e1-7.

333. Hershman DL, Unger JM, Barlow WE, Hutchins LF, Martino S, Osborne CK,

Livingston RB, Albain KS. Treatment quality and outcomes of African American versus

white breast cancer patients: retrospective analysis of Southwest Oncology studies

S8814/S8897. J Clin Oncol. 2009;27(13):2157-62.

334. Griggs JJ, Culakova E, Sorbero ME, van Ryn M, Poniewierski MS, Wolff DA,

Crawford J, Dale DC, Lyman GH. Effect of patient socioeconomic status and body mass

index on the quality of breast cancer adjuvant chemotherapy. J Clin Oncol.

2007;25(3):277-84.

335. Freedman RA, Virgo KS, He Y, Pavluck AL, Winer EP, Ward EM, Keating NL. The

association of race/ethnicity, insurance status, and socioeconomic factors with breast

cancer care. Cancer. 2011;117(1):180-9.

336. Goldhirsch A, Ingle JN, Gelber R, Coates A, Thürlimann B, Senn H-J. Thresholds for

therapies: highlights of the St Gallen International Expert Consensus on the primary

therapy of early breast cancer 2009. Ann Oncol. 2009;20(8):1319-29.

337. Ayanian JZ, Zaslavsky AM, Fuchs CS, Guadagnoli E, Creech CM, Cress RD, O’Connor

LC, West DW, Allen ME, Wolf RE. Use of adjuvant chemotherapy and radiation

therapy for colorectal cancer in a population-based cohort. J Clin Oncol.

2003;21(7):1293-300.

338. Griggs JJ, Sorbero ME, Stark AT, Heininger SE, Dick AW. Racial disparity in the dose

and dose intensity of breast cancer adjuvant chemotherapy. Breast Cancer Res Treat.

2003;81(1):21-31.

339. Smith K, Wray L, Klein-Cabral M, Schuchter L, Fox K, Glick J, DeMichele A. Ethnic

disparities in adjuvant chemotherapy for breast cancer are not caused by excess toxicity

in black patients. Clin Breast Cancer. 2005;6(3):260-6.

340. Schroen AT, Brenin DR, Kelly MD, Knaus WA, Slingluff CL. Impact of patient

distance to radiation therapy on mastectomy use in early-stage breast cancer patients. J

Clin Oncol. 2005;23(28):7074-80.

341. Gilligan MA, Kneusel RT, Hoffmann RG, Greer AL, Nattinger AB. Persistent

differences in sociodemographic determinants of breast conserving treatment despite

overall increased adoption. Med Care. 2002;40(3):181-9.

342. Davies C, Godwin J, Gray R, Clarke M, Cutter D, Darby S, McGale P, Pan HC, Taylor

C, Wang YC, Dowsett M, Ingle J, Peto R. Relevance of breast cancer hormone receptors

and other factors to the efficacy of adjuvant tamoxifen: patient-level meta-analysis of

randomised trials. Lancet. 2011;378(9793):771-84.

References

301

343. Ruddy K, Mayer E, Partridge A. Patient adherence and persistence with oral anticancer

treatment. CA Cancer J Clin. 2009;59(1):56-66.

344. Makubate B, Donnan PT, Dewar JA, Thompson AM, McCowan C. Cohort study of

adherence to adjuvant endocrine therapy, breast cancer recurrence and mortality. Br J

Cancer. 2013;108(7):1515-24.

345. Barron TI, Connolly R, Bennett K, Feely J, Kennedy MJ. Early discontinuation of

tamoxifen: a lesson for oncologists. Cancer. 2007;109(5):832-9.

346. Ma AM, Barone J, Wallis AE, Wu NJ, Garcia LB, Estabrook A, Rosenbaum-Smith SM,

Tartter PI. Noncompliance with adjuvant radiation, chemotherapy, or hormonal therapy

in breast cancer patients. Am J Surg. 2008;196(4):500-4.

347. McCowan C, Shearer J, Donnan PT, Dewar JA, Crilly M, Thompson AM, Fahey TP.

Cohort study examining tamoxifen adherence and its relationship to mortality in women

with breast cancer. Br J Cancer. 2008;99(11):1763-8.

348. Forbes JF, Cuzick J, Buzdar A, Howell A, Tobias JS, Baum M. Effect of anastrozole and

tamoxifen as adjuvant treatment for early-stage breast cancer: 100-month analysis of the

ATAC trial. Lancet Oncol. 2008;9(1):45-53.

349. Jatrana S, Crampton P. Primary health care in New Zealand: who has access? Health

Policy. 2009;93(1):1-10.

350. Jatrana S, Crampton P, Norris P. Ethnic differences in access to prescription medication

because of cost in New Zealand. J Epidemiol Community Health. 2011;65(5):454-60.

351. Jacob Arriola KR, Mason TA, Bannon KA, Holmes C, Powell CL, Horne K, O’Regan

R. Modifiable risk factors for adherence to adjuvant endocrine therapy among breast

cancer patients. Patient Educ Couns. 2014;95(1):98-103.

352. Koay K, Schofield P, Jefford M. Importance of health literacy in oncology. Asia Pac J

Clin Oncol. 2012;8(1):14-23.

353. NIH consensus conference. Treatment of early-stage breast cancer: guidelines. JAMA.

1991;265(3):391-5.

354. Krag DN, Anderson SJ, Julian TB, Brown AM, Harlow SP, Costantino JP, Ashikaga T,

Weaver DL, Mamounas EP, Jalovec LM. Sentinel-lymph-node resection compared with

conventional axillary-lymph-node dissection in clinically node-negative patients with

breast cancer: overall survival findings from the NSABP B-32 randomised phase 3 trial.

Lancet Oncol. 2010;11(10):927-33.

355. National Breast and Ovarian Cancer Centre (NBOCC). Recommendations for use of

Sentinel node biopsy in early (operable) breast cancer. Australia: Cancer Australia;

2011.

356. Lyman GH, Giuliano AE, Somerfield MR, Benson AB, 3rd, Bodurka DC, Burstein HJ,

Cochran AJ, Cody HS, 3rd, Edge SB, Galper S, Hayman JA, Kim TY, Perkins CL,

Podoloff DA, Sivasubramaniam VH, Turner RR, Wahl R, Weaver DL, Wolff AC, Winer

EP. American Society of Clinical Oncology guideline recommendations for sentinel

lymph node biopsy in early-stage breast cancer. J Clin Oncol. 2005;23(30):7703-20.


302

357. Nesvold I-L, Dahl AA, Løkkevik E, Marit Mengshoel A, Fosså SD. Arm and shoulder

morbidity in breast cancer patients after breast-conserving therapy versus mastectomy.

Acta Oncol. 2008;47(5):835-42.

358. Burak Jr WE, Hollenbeck ST, Zervos EE, Hock KL, Kemp LC, Young DC. Sentinel

lymph node biopsy results in less postoperative morbidity compared with axillary lymph

node dissection for breast cancer. Am J Surg. 2002;183(1):23-7.

359. Albornoz CR, Bach PB, Mehrara BJ, Disa JJ, Pusic AL, McCarthy CM, Cordeiro PG,

Matros E. A paradigm shift in US Breast reconstruction: increasing implant rates. Plast

Reconstr Surg. 2013;131(1):15-23.

360. Nold RJ, Beamer RL, Helmer SD, McBoyle MF. Factors influencing a woman's choice

to undergo breast-conserving surgery versus modified radical mastectomy. Am J Surg.

2000;180(6):413-8.

361. Morrow M, White J, Moughan J, Owen J, Pajack T, Sylvester J, Wilson JF, Winchester

D. Factors predicting the use of breast-conserving therapy in stage I and II breast

carcinoma. J Clin Oncol. 2001;19(8):2254-62.

362. Brennan M, Spillane A. Uptake and predictors of post-mastectomy reconstruction in

women with breast malignancy: systematic review. Eur J Surg Oncol. 2013;39(6):527-

41.

363. Roder D, Zorbas H, Kollias J, Pyke C, Walters D, Campbell I, Taylor C, Webster F.

Factors predictive of immediate breast reconstruction following mastectomy for invasive

breast cancer in Australia. Breast. 2013;22(6):1220-5.

364. Hislop TG, Olivotto IA, Coldman AJ, Trevisan CH, Kula J, McGregor GI, Phillips N.

Variations in breast conservation surgery for women with axillary lymph node negative

breast cancer in British Columbia. Can J Public Health. 1996;87(6):390-4.

365. Jeevan R, Cromwell DA, Browne JP, Trivella M, Pereira J, Caddy CM, Sheppard C, van

der Meulen JH. Regional variation in use of immediate breast reconstruction after

mastectomy for breast cancer in England. Eur J Surg Oncol. 2010;36(8):750-5.

366. Tseng JF, Kronowitz SJ, Sun CC, Perry AC, Hunt KK, Babiera GV, Newman LA,

Singletary SE, Mirza NQ, Ames FC, Meric-Bernstam F, Ross MI, Feig BW, Robb GL,

Kuerer HM. The effect of ethnicity on immediate reconstruction rates after mastectomy

for breast cancer. Cancer. 2004;101(7):1514-23.

367. Caldon L, Collins K, Wilde D, Ahmedzai S, Noble T, Stotter A, Sibbering D, Holt S,

Reed M. Why do hospital mastectomy rates vary: Differences in the decision-making

experiences of women with breast cancer. Br J Cancer. 2011;104(10):1551-7.

368. Gaudette LA, Gao R-N, Spence A, Shi F, Johansen H, Olivotto IA. Declining use of

mastectomy for invasive breast cancer in Canada, 1981-2000. Can J Public Health.

2004:336-40.

369. Mastaglia B, Kristjanson LJ. Factors influencing women's decisions for choice of

surgery for Stage I and Stage II breast cancer in Western Australia. J Adv Nurs.

2001;35(6):836-47.

References

303

370. Fallowfield LJ, Baum M, Maguire GP. Effects of breast conservation on psychological

morbidity associated with diagnosis and treatment of early breast cancer. British medical

journal (Clinical research ed). 1986;293(6558):1331-4.

371. Benedict S, Cole DJ, Baron L, Baron P. Factors influencing choice between mastectomy

and lumpectomy for women in the Carolinas. J Surg Oncol. 2001;76(1):6-12.

372. Pham JT, Allen LJ, Gomez SL. Why do Asian-American women have lower rates of

breast conserving surgery: results of a survey regarding physician perceptions. BMC

Public Health. 2009;9:246.

373. Minicozzi P, Cirilli C, Federico M, Capocaccia R, Budroni M, Candela P, Falcini F,

Fusco M, Giacomin A, La Rosa F. Differences in stage and treatment of breast cancer

across Italy point to inequalities in access to and availability of proper care. Tumori.

2012;98(2):204-9.

374. Morrow M, Mujahid M, Lantz PM, Janz NK, Fagerlin A, Schwartz K, Liu L, Deapen D,

Salem B, Lakhani I. Correlates of breast reconstruction. Cancer. 2005;104(11):2340-6.

375. Bach PB, Schrag D, Brawley OW, Galaznik A, Yakren S, Begg CB. Survival of blacks

and whites after a cancer diagnosis. JAMA. 2002;287(16):2106-13.

376. Marmot M, Friel S, Bell R, Houweling TA, Taylor S. Closing the gap in a generation:

health equity through action on the social determinants of health. Lancet.

2008;372(9650):1661-9.

377. Lund MJ, Trivers KF, Porter PL, Coates RJ, Leyland-Jones B, Brawley OW, Flagg EW,

O’Regan RM, Gabram SG, Eley JW. Race and triple negative threats to breast cancer

survival: a population-based study in Atlanta, GA. Breast Cancer Res Treat.

2009;113(2):357-70.

378. Mitchell EA. Racial inequalities in childhood asthma. Soc Sci Med. 1991;32(7):831-6.

379. New Zealand Health Information Service. Cancer patient survival : change over time

update: covering the period 1994-2009. Wellington2012. p. v.

380. Jemal A, Siegel R, Xu J, Ward E. Cancer statistics, 2010. CA Cancer J Clin.

2010;60(5):277-300.

381. Whitehead M. The concepts and principles of equity and health. Int J Health Serv.

1992;22(3):429-45.

382. Ministry of Health. Reducing inequalities in health [Internet]Wellington2002 [Available

from: http://www.health.govt.nz/publication/reducing-inequalities-health

383. Carmichael AR. Obesity and prognosis of breast cancer. Obesity reviews : an official

journal of the International Association for the Study of Obesity. 2006;7(4):333-40.

384. Mackenbach JP, Bakker M. Reducing inequalities in health: a European perspective.

London ; New York: Routledge; 2002.

385. Seneviratne S, Campbell I, Scott N, Coles C, Lawrenson R. Treatment delay for Māori

women with breast cancer in New Zealand. Ethnicity & health. 2015;20(2):178-93.

http://www.health.govt.nz/publication/reducing-inequalities-health


304

386. Craig P, Dieppe P, Macintyre S, Michie S, Nazareth I, Petticrew M. Developing and

evaluating complex interventions: new guidance. London, UK: Medical Research

Council, UK; 2008.

387. Scott N. Information and recommendations to assist the Ministry of Health when

prioritising Māori Cancer Control issues: Based on Evidence from Unequal Impact II:

Māori and Non-Māori Cancer Statistics by Deprivation and Rural Urban Status, 2002 –

2006. A report commissioned by Te Kete Hauora, Ministry of Health; 2011

(Unpublished report).

388. Bonfill Cosp X, Marzo Castillejo M, Pladevall Vila M, Marti J, Emparanza JI. Strategies

for increasing the participation of women in community breast cancer screening.

Cochrane Database Syst Rev. 2001;1.

389. Vogel BA, Bengel J, Helmes AW. Information and decision making: patients’ needs and

experiences in the course of breast cancer treatment. Patient Educ Couns. 2008;71(1):79-

85.

390. Yates P. Cancer care coordinators: realising the potential for improving the patient

journey. Cancer Forum. 2004;28(3):128.

391. Lawrenson R, Smyth D, Kara E, Thomson R. Rural general practitioner perspectives of

the needs of Māori patients requiring palliative care. N Z Med J. 2010;123(1315).

392. Kenfield SA, Stampfer MJ, Rosner BA, Colditz GA. Smoking and smoking cessation in

relation to mortality in women. JAMA. 2008;299(17):2037-47.

393. Jackson R, Graham P, Beaglehole R, De Boer G. Validation of coronary heart disease

death certificate diagnoses. N Z Med J. 1988;101(856 Pt 1):658-60.

394. Ajwani S. Unlocking the numerator-denominator bias for the 1980s and 1990s

[Internet]. [Rev ed. Wellington: Dept. of Public Health, School of Medicine and Health

Sciences; 2003 [Available from: http://ndhadeliver.natlib.govt.nz/content-

aggregator/getIEs?system=ilsdb&id=1397008

395. Robson B, Reid P. Ethnicity matters review of the measurement of ethnicity in official

statistics: Māori perspectives paper for consultation [Internet]Wellington: Statistics New

Zealand; 2001 [Available from:

http://www.stats.govt.nz/browse_for_stats/population/census_counts/review-

measurement-of-ethnicity/Māori-perspective.aspx

396. Cormack BE, Purdie G, Robson B. Cancer. In: Robson B, Harris R, editors. Hauora,

Māori standards of health IV: a study of the years, 2000-2005. Wellington: Te Ropu

Rangahau Hauora a Eru Pomare; 2007. p. xii, 273 p.

397. Salmond CE, Crampton P. Development of New Zealand's deprivation index (NZDep)

and its uptake as a national policy tool. Can J Public Health. 2012;103(8 Suppl 2):S7-11.

398. Jansen P. Non-financial barriers to primary health care services for Māori. J Prim Health

Care. 2009;1(3):240.

http://ndhadeliver.natlib.govt.nz/content-aggregator/getIEs?system=ilsdb&id=1397008

http://ndhadeliver.natlib.govt.nz/content-aggregator/getIEs?system=ilsdb&id=1397008

http://www.stats.govt.nz/browse_for_stats/population/census_counts/review-measurement-of-ethnicity/maori-perspective.aspx

http://www.stats.govt.nz/browse_for_stats/population/census_counts/review-measurement-of-ethnicity/maori-perspective.aspx

References

305

Documents

researchspace.auckland.ac - Semantic Scholar...Dr. Sanjeewa Anuruddha Seneviratne A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy in Surgery,